Showing 1–50 of 61 results for author: Chowdhury, F

Simulating spin biology using a digital quantum computer: Prospects on a near-term quantum hardware emulator

Authors: Pedro H. Alvarez, Farhan T. Chowdhury, Luke D. Smith, Trevor J. Brokowski, Clarice D. Aiello, Daniel R. Kattnig, Marcos C. de Oliveira

Abstract: Understanding the intricate quantum spin dynamics of radical pair reactions is crucial for unraveling the underlying nature of chemical processes across diverse scientific domains. In this work, we leverage Trotterization to map coherent radical pair spin dynamics onto a digital gate-based quantum simulation. Our results demonstrated agreement between the idealized noiseless quantum circuit simula… ▽ More Understanding the intricate quantum spin dynamics of radical pair reactions is crucial for unraveling the underlying nature of chemical processes across diverse scientific domains. In this work, we leverage Trotterization to map coherent radical pair spin dynamics onto a digital gate-based quantum simulation. Our results demonstrated agreement between the idealized noiseless quantum circuit simulation and established master equation approaches for homogeneous radical pair recombination, identifying approximately 15 Trotter steps to be sufficient for faithfully reproducing the coupled spin dynamics of a prototypical system. By utilizing this computational technique to study the dynamics of spin systems of biological relevance, our findings underscore the potential of digital quantum simulation (DQS) of complex radical pair reactions and builds the groundwork towards more utilitarian investigations into their intricate reaction dynamics. We further investigate the effect of realistic error models on our DQS approach, and provide an upper limit for the number of Trotter steps that can currently be applied in the absence of error mitigation techniques before losing simulation accuracy to deleterious noise effects. △ Less

Submitted 18 June, 2024; originally announced June 2024.

Comments: 10 pages, 9 figures

Anti-seizure medication tapering is associated with delta band power reduction in a dose, region and time-dependent manner

Authors: Guillermo M. Besne, Nathan Evans, Mariella Panagiotopoulou, Billy Smith, Fahmida A Chowdhury, Beate Diehl, John S Duncan, Andrew W McEvoy, Anna Miserocchi, Jane de Tisi, Mathew Walker, Peter N. Taylor, Chris Thornton, Yujiang Wang

Abstract: Anti-seizure medications (ASMs) are the primary treatment for epilepsy, yet medication tapering effects have not been investigated in a dose, region, and time-dependent manner, despite their potential impact on research and clinical practice. We examined over 3000 hours of intracranial EEG recordings in 32 subjects during long-term monitoring, of which 22 underwent concurrent ASM tapering. We es… ▽ More Anti-seizure medications (ASMs) are the primary treatment for epilepsy, yet medication tapering effects have not been investigated in a dose, region, and time-dependent manner, despite their potential impact on research and clinical practice. We examined over 3000 hours of intracranial EEG recordings in 32 subjects during long-term monitoring, of which 22 underwent concurrent ASM tapering. We estimated ASM plasma levels based on known pharmaco-kinetics of all the major ASM types. We found an overall decrease in the power of delta band activity around the period of maximum medication withdrawal in most (80%) subjects, independent of their epilepsy type or medication combination. The degree of withdrawal correlated positively with the magnitude of delta power decrease. This dose-dependent effect was strongly seen across all recorded cortical regions during daytime; but not in sub-cortical regions, or during night time. We found no evidence of differential effect in seizure onset, spiking, or pathological brain regions. The finding of decreased delta band power during ASM tapering agrees with previous literature. Our observed dose-dependent effect indicates that monitoring ASM levels in cortical regions may be feasible for applications such as medication reminder systems, or closed-loop ASM delivery systems. ASMs are also used in other neurological and psychiatric conditions, making our findings relevant to a general neuroscience and neurology audience. △ Less

Submitted 2 May, 2024; originally announced May 2024.

CAPE: CAM as a Probabilistic Ensemble for Enhanced DNN Interpretation

Authors: Townim Faisal Chowdhury, Kewen Liao, Vu Minh Hieu Phan, Minh-Son To, Yutong Xie, Kevin Hung, David Ross, Anton van den Hengel, Johan W. Verjans, Zhibin Liao

Abstract: Deep Neural Networks (DNNs) are widely used for visual classification tasks, but their complex computation process and black-box nature hinder decision transparency and interpretability. Class activation maps (CAMs) and recent variants provide ways to visually explain the DNN decision-making process by displaying 'attention' heatmaps of the DNNs. Nevertheless, the CAM explanation only offers relat… ▽ More Deep Neural Networks (DNNs) are widely used for visual classification tasks, but their complex computation process and black-box nature hinder decision transparency and interpretability. Class activation maps (CAMs) and recent variants provide ways to visually explain the DNN decision-making process by displaying 'attention' heatmaps of the DNNs. Nevertheless, the CAM explanation only offers relative attention information, that is, on an attention heatmap, we can interpret which image region is more or less important than the others. However, these regions cannot be meaningfully compared across classes, and the contribution of each region to the model's class prediction is not revealed. To address these challenges that ultimately lead to better DNN Interpretation, in this paper, we propose CAPE, a novel reformulation of CAM that provides a unified and probabilistically meaningful assessment of the contributions of image regions. We quantitatively and qualitatively compare CAPE with state-of-the-art CAM methods on CUB and ImageNet benchmark datasets to demonstrate enhanced interpretability. We also test on a cytology imaging dataset depicting a challenging Chronic Myelomonocytic Leukemia (CMML) diagnosis problem. Code is available at: https://github.com/AIML-MED/CAPE. △ Less

Submitted 4 April, 2024; v1 submitted 2 April, 2024; originally announced April 2024.

On the optimality of the radical-pair quantum compass

Authors: Luke D. Smith, Jonas Glatthard, Farhan T. Chowdhury, Daniel R. Kattnig

Abstract: Quantum sensing enables the ultimate precision attainable in parameter estimation. Circumstantial evidence suggests that certain organisms, most notably migratory songbirds, also harness quantum-enhanced magnetic field sensing via a radical-pair-based chemical compass for the precise detection of the weak geomagnetic field. However, what underpins the acuity of such a compass operating in a noisy… ▽ More Quantum sensing enables the ultimate precision attainable in parameter estimation. Circumstantial evidence suggests that certain organisms, most notably migratory songbirds, also harness quantum-enhanced magnetic field sensing via a radical-pair-based chemical compass for the precise detection of the weak geomagnetic field. However, what underpins the acuity of such a compass operating in a noisy biological setting, at physiological temperatures, remains an open question. Here, we address the fundamental limits of inferring geomagnetic field directions from radical-pair spin dynamics. Specifically, we compare the compass precision, as derived from the directional dependence of the radical-pair recombination yield, to the ultimate precision potentially realisable by a quantum measurement on the spin system under steady-state conditions. To this end, we probe the quantum Fisher information and associated Cramér--Rao bound in spin models of realistic complexity, accounting for complex inter-radical interactions, a multitude of hyperfine couplings, and asymmetric recombination kinetics, as characteristic for the magnetosensory protein cryptochrome. We compare several models implicated in cryptochrome magnetoreception and unveil their optimality through the precision of measurements ostensibly accessible to nature. Overall, the comparison provides insight into processes honed by nature to realise optimality whilst constrained to operating with mere reaction yields. Generally, the inference of compass orientation from recombination yields approaches optimality in the limits of complexity, yet plateaus short of the theoretical optimal precision bounds by up to one or two orders of magnitude, thus underscoring the potential for improving on design principles inherent to natural systems. △ Less

Submitted 5 January, 2024; originally announced January 2024.

Journal ref: Quantum Science and Technology 9, 035041 (2024)

Proximal quantum control of spin and spin ensemble with highly localized control field from skyrmions

Authors: Md Fahim F Chowdhury, Mohamad Niknam, Md Mahadi Rajib, Louis S. Bouchard, Jayasimha Atulasimha

Abstract: Selective control of individual spin qubits is needed for scalable quantum computing based on spin states. Achieving high-fidelity in both single and two-qubit gates, essential components of universal quantum computers, necessitates highly localized control fields. These fields must be capable of addressing specific spin qubits while minimizing gate errors and cross-talk in adjacent qubits. Overco… ▽ More Selective control of individual spin qubits is needed for scalable quantum computing based on spin states. Achieving high-fidelity in both single and two-qubit gates, essential components of universal quantum computers, necessitates highly localized control fields. These fields must be capable of addressing specific spin qubits while minimizing gate errors and cross-talk in adjacent qubits. Overcoming the challenge of generating a localized radio-frequency magnetic field, in the absence of elementary magnetic monopoles, we introduce a technique that combines divergent and convergent nanoscale magnetic skyrmions. This approach produces a precise control field that manipulates spin qubits with high fidelity. We propose the use of 2D skyrmions, which are 2D analogues of 3D hedgehog structures. The latter are emergent magnetic monopoles, but difficult to fabricate. The 2D skyrmions, on the other hand, can be fabricated using standard semiconductor foundry processes. Our comparative analysis of the density matrix evolution and gate fidelities in scenarios involving proximal skyrmions and nanomagnets indicates potential gate fidelities surpassing 99.95% for π/2-gates and 99.90% for π-gates. Notably, the skyrmion configuration generates a significantly lower field on neighboring spin qubits, i.e. 15 times smaller field on a neighboring qubit compared to nanomagnets that produces the same field at the controlled qubit, making it a more suitable candidate for scalable quantum control architectures by reducing disturbances in adjacent qubits. △ Less

Submitted 31 December, 2023; originally announced January 2024.

Incomplete resection of the icEEG seizure onset zone is not associated with post-surgical outcomes

Authors: Sarah J. Gascoigne, Nathan Evans, Gerard Hall, Csaba Kozma, Mariella Panagiotopoulou, Gabrielle M. Schroeder, Callum Simpson, Christopher Thornton, Frances Turner, Heather Woodhouse, Jess Blickwedel, Fahmida Chowdhury, Beate Diehl, John S. Duncan, Ryan Faulder, Rhys H. Thomas, Kevin Wilson, Peter N. Taylor, Yujiang Wang

Abstract: Delineation of seizure onset regions from EEG is important for effective surgical workup. However, it is unknown if their complete resection is required for seizure freedom, or in other words, if post-surgical seizure recurrence is due to incomplete removal of the seizure onset regions. Retrospective analysis of icEEG recordings from 63 subjects (735 seizures) identified seizure onset regions th… ▽ More Delineation of seizure onset regions from EEG is important for effective surgical workup. However, it is unknown if their complete resection is required for seizure freedom, or in other words, if post-surgical seizure recurrence is due to incomplete removal of the seizure onset regions. Retrospective analysis of icEEG recordings from 63 subjects (735 seizures) identified seizure onset regions through visual inspection and algorithmic delineation. We analysed resection of onset regions and correlated this with post-surgical seizure control. Most subjects had over half of onset regions resected (70.7% and 60.5% of subjects for visual and algorithmic methods, respectively). In investigating spatial extent of onset or resection, and presence of diffuse onsets, we found no substantial evidence of association with post-surgical seizure control (all AUC<0.7, p>0.05). Seizure onset regions tends to be at least partially resected, however a less complete resection is not associated with worse post-surgical outcome. We conclude that seizure recurrence after epilepsy surgery is not necessarily a result of failing to completely resect the seizure onset zone, as defined by icEEG. Other network mechanisms must be involved, which are not limited to seizure onset regions alone. △ Less

Submitted 24 November, 2023; originally announced November 2023.

ChatGPT-guided Semantics for Zero-shot Learning

Authors: Fahimul Hoque Shubho, Townim Faisal Chowdhury, Ali Cheraghian, Morteza Saberi, Nabeel Mohammed, Shafin Rahman

Abstract: Zero-shot learning (ZSL) aims to classify objects that are not observed or seen during training. It relies on class semantic description to transfer knowledge from the seen classes to the unseen classes. Existing methods of obtaining class semantics include manual attributes or automatic word vectors from language models (like word2vec). We know attribute annotation is costly, whereas automatic wo… ▽ More Zero-shot learning (ZSL) aims to classify objects that are not observed or seen during training. It relies on class semantic description to transfer knowledge from the seen classes to the unseen classes. Existing methods of obtaining class semantics include manual attributes or automatic word vectors from language models (like word2vec). We know attribute annotation is costly, whereas automatic word-vectors are relatively noisy. To address this problem, we explore how ChatGPT, a large language model, can enhance class semantics for ZSL tasks. ChatGPT can be a helpful source to obtain text descriptions for each class containing related attributes and semantics. We use the word2vec model to get a word vector using the texts from ChatGPT. Then, we enrich word vectors by combining the word embeddings from class names and descriptions generated by ChatGPT. More specifically, we leverage ChatGPT to provide extra supervision for the class description, eventually benefiting ZSL models. We evaluate our approach on various 2D image (CUB and AwA) and 3D point cloud (ModelNet10, ModelNet40, and ScanObjectNN) datasets and show that it improves ZSL performance. Our work contributes to the ZSL literature by applying ChatGPT for class semantics enhancement and proposing a novel word vector fusion method. △ Less

Submitted 17 October, 2023; originally announced October 2023.

Comments: Accepted in International Conference on Digital Image Computing: Techniques and Applications (DICTA), 2023

Diminished circadian and ultradian rhythms in pathological brain tissue in human in vivo

Authors: Mariella Panagiotopoulou, Christopher Thornton, Fahmida A Chowdhury, Beate Diehl, John S Duncan, Sarah J Gascoigne, Andrew W McEvoy, Anna Miserocchi, Billy C Smith, Jane de Tisi, Peter N Taylor, Yujiang Wang

Abstract: Chronobiological rhythms, such as the circadian rhythm, have long been linked to neurological disorders, but it is currently unknown how pathological processes affect the expression of biological rhythms in the brain. Here, we use the unique opportunity of long-term, continuous intracranially recorded EEG from 38 patients (totalling 6338 hours) to delineate circadian and ultradian rhythms in diffe… ▽ More Chronobiological rhythms, such as the circadian rhythm, have long been linked to neurological disorders, but it is currently unknown how pathological processes affect the expression of biological rhythms in the brain. Here, we use the unique opportunity of long-term, continuous intracranially recorded EEG from 38 patients (totalling 6338 hours) to delineate circadian and ultradian rhythms in different brain regions. We show that functional circadian and ultradian rhythms are diminished in pathological tissue, independent of regional variations. We further demonstrate that these diminished rhythms are persistent in time, regardless of load or occurrence of pathological events. These findings provide the first evidence that brain pathology is functionally associated with persistently diminished chronobiological rhythms in vivo in humans, independent of regional variations or pathological events. Future work interacting with, and restoring, these modulatory chronobiological rhythms may allow for novel therapies. △ Less

Submitted 13 September, 2023; originally announced September 2023.

Quantum Control of Radical Pair Dynamics beyond Time-Local Optimization

Authors: Farhan T. Chowdhury, Matt C. J. Denton, Daniel C. Bonser, Daniel R. Kattnig

Abstract: We realize arbitrary waveform-based control of spin-selective recombination reactions of radical pairs in the low magnetic field regime. To this end, we extend the Gradient Ascent Pulse Engineering (GRAPE) paradigm to allow for optimizing reaction yields. This overcomes drawbacks of previously suggested time-local optimization approaches for the reaction control of radical pairs, which were limite… ▽ More We realize arbitrary waveform-based control of spin-selective recombination reactions of radical pairs in the low magnetic field regime. To this end, we extend the Gradient Ascent Pulse Engineering (GRAPE) paradigm to allow for optimizing reaction yields. This overcomes drawbacks of previously suggested time-local optimization approaches for the reaction control of radical pairs, which were limited to high biasing fields. We demonstrate how efficient time-global optimization of the recombination yields can be realized by gradient based methods augmented by time-blocking, sparse sampling of the yield, and evaluation of the central single time-step propagators and their Fréchet derivatives using iterated Trotter-Suzuki splittings. Results are shown for both a toy model, previously used to demonstrate coherent control of radical pair reactions in the simpler high-field scenario, and furthermore for a realistic exciplex-forming donor-acceptor system comprising 16 nuclear spins. This raises prospects for the spin-control of actual radical pair systems in ambient magnetic fields, by suppressing or boosting radical reaction yields using purpose-specific radio-frequency waveforms, paving the way for reaction-yield-dependent quantum magnetometry and potentially applications of quantum control to biochemical radical pair reactions. We demonstrate the latter aspect for two radical pairs implicated in quantum biology. △ Less

Submitted 13 March, 2024; v1 submitted 14 June, 2023; originally announced June 2023.

Comments: 17 pages, 9 figures, plus supplementary material (8 pagese, 8 figures)

Journal ref: PRX Quantum 5, 020303 (2024)

Interictal MEG abnormalities to guide intracranial electrode implantation and predict surgical outcome

Authors: Thomas W. Owen, Vytene Janiukstyte, Gerard R. Hall, Fahmida A. Chowdhury, Beate Diehl, Andrew McEvoy, Anna Miserocchi, Jane de Tisi, John S. Duncan, Fergus Rugg-Gunn, Yujiang Wang, Peter N. Taylor

Abstract: Intracranial EEG (iEEG) is the gold standard technique for epileptogenic zone (EZ) localisation, but requires a hypothesis of which tissue is epileptogenic, guided by qualitative analysis of seizure semiology and other imaging modalities such as magnetoencephalography (MEG). We hypothesised that if quantifiable MEG band power abnormalities were sampled by iEEG, then patients' post-resection seizur… ▽ More Intracranial EEG (iEEG) is the gold standard technique for epileptogenic zone (EZ) localisation, but requires a hypothesis of which tissue is epileptogenic, guided by qualitative analysis of seizure semiology and other imaging modalities such as magnetoencephalography (MEG). We hypothesised that if quantifiable MEG band power abnormalities were sampled by iEEG, then patients' post-resection seizure outcome were better. Thirty-two individuals with neocortical epilepsy underwent MEG and iEEG recordings as part of pre-surgical evaluation. Interictal MEG band power abnormalities were derived using 70 healthy controls as a normative baseline. MEG abnormality maps were compared to electrode implantation, with the spatial overlap of iEEG electrodes and MEG abnormalities recorded. Finally, we assessed if the implantation of electrodes in abnormal tissue, and resection of the strongest abnormalities determined by MEG and iEEG explained surgical outcome. Intracranial electrodes were implanted in brain tissue with the most abnormal MEG findings in individuals that were seizure-free post-resection (T=3.9, p=0.003). The overlap between MEG abnormalities and iEEG electrodes distinguished outcome groups moderately well (AUC=0.68). In isolation, the resection of the strongest MEG and iEEG abnormalities separated surgical outcome groups well (AUC=0.71, AUC=0.74 respectively). A model incorporating all three features separated outcome groups best (AUC=0.80). Intracranial EEG is a key tool to delineate the EZ and help render patients seizure-free after resection. We showed that data-driven abnormalities derived from interictal MEG recordings have clinical value and may help guide electrode placement in individuals with neocortical epilepsy. Finally, our predictive model of post-operative seizure-freedom, which leverages both MEG and iEEG recordings, may aid patient counselling of expected outcome. △ Less

Submitted 11 April, 2023; originally announced April 2023.

Comments: 22 pages, 6 figures

Complementary structural and functional abnormalities to localise epileptogenic tissue

Authors: Jonathan J Horsley, Rhys H Thomas, Fahmida A Chowdhury, Beate Diehl, Andrew W McEvoy, Anna Miserocchi, Jane de Tisi, Sjoerd B Vos, Matthew C Walker, Gavin P Winston, John S Duncan, Yujiang Wang, Peter N Taylor

Abstract: When investigating suitability for surgery, people with drug-refractory focal epilepsy may have intracranial EEG (iEEG) electrodes implanted to localise seizure onset. Diffusion-weighted magnetic resonance imaging (dMRI) may be acquired to identify key white matter tracts for surgical avoidance. Here, we investigate whether structural connectivity abnormalities, inferred from dMRI, may be used in… ▽ More When investigating suitability for surgery, people with drug-refractory focal epilepsy may have intracranial EEG (iEEG) electrodes implanted to localise seizure onset. Diffusion-weighted magnetic resonance imaging (dMRI) may be acquired to identify key white matter tracts for surgical avoidance. Here, we investigate whether structural connectivity abnormalities, inferred from dMRI, may be used in conjunction with functional iEEG abnormalities to aid localisation and resection of the epileptogenic zone (EZ), and improve surgical outcomes in epilepsy. We retrospectively investigated data from 43 patients with epilepsy who had surgery following iEEG. Twenty five patients (58%) were free from disabling seizures (ILAE 1 or 2) at one year. For all patients, T1-weighted and diffusion-weighted MRIs were acquired prior to iEEG implantation. Interictal iEEG functional, and dMRI structural connectivity abnormalities were quantified by comparison to a normative map and healthy controls respectively. First, we explored whether the resection of maximal (dMRI and iEEG) abnormalities related to improved surgical outcomes. Second, we investigated whether the modalities provided complementary information for improved prediction of surgical outcome. Third, we suggest how dMRI abnormalities may be useful to inform the placement of iEEG electrodes as part of the pre-surgical evaluation using a patient case study. Seizure freedom was 15 times more likely in those patients with resection of maximal dMRI and iEEG abnormalities (p=0.008). Both modalities were separately able to distinguish patient outcome groups and when combined, a decision tree correctly separated 36 out of 43 (84%) patients based on surgical outcome. Structural dMRI could be used in pre-surgical evaluations, particularly when localisation of the EZ is uncertain, to inform personalised iEEG implantation and resection. △ Less

Submitted 24 October, 2023; v1 submitted 6 April, 2023; originally announced April 2023.

Comments: 5 figures

Temporal stability of intracranial EEG abnormality maps for localising epileptogenic tissue

Authors: Yujiang Wang, Gabrielle M Schroeder, Jonathan J Horsley, Mariella Panagiotopoulou, Fahmida A Chowdhury, Beate Diehl, John S Duncan, Andrew W McEvoy, Anna Miserocchi, Jane de Tisi, Peter N Taylor

Abstract: Objective: Identifying abnormalities in interictal intracranial EEG, by comparing patient data to a normative map, has shown promise for the localisation of epileptogenic tissue and prediction of outcome. The approach typically uses short interictal segments of around one minute. However, the temporal stability of findings has not been established. Methods: Here, we generated a normative map of… ▽ More Objective: Identifying abnormalities in interictal intracranial EEG, by comparing patient data to a normative map, has shown promise for the localisation of epileptogenic tissue and prediction of outcome. The approach typically uses short interictal segments of around one minute. However, the temporal stability of findings has not been established. Methods: Here, we generated a normative map of iEEG in non-pathological brain tissue from 249 patients. We computed regional band power abnormalities in a separate cohort of 39 patients for the duration of their monitoring period (0.92-8.62 days of iEEG data, mean 4.58 days per patient, over 4,800 hours recording). To assess the localising value of band power abnormality, we computed DRS - a measure of how different the surgically resected and spared tissue were in terms of band power abnormalities - over time. Results: In each patient, band power abnormality was relatively consistent over time. The median DRS of the entire recording period separated seizure free (ILAE = 1) and not seizure free (ILAE > 1) patients well (AUC = 0.69). This effect was similar interictally (AUC = 0.69) and peri-ictally (AUC = 0.71). Significance: Our results suggest that band power abnormality DRS, as a predictor of outcomes from epilepsy surgery, is a relatively robust metric over time. These findings add further support for abnormality map** of neurophysiology data during presurgical evaluation. △ Less

Submitted 11 February, 2023; originally announced February 2023.

KnowGL: Knowledge Generation and Linking from Text

Authors: Gaetano Rossiello, Md Faisal Mahbub Chowdhury, Nandana Mihindukulasooriya, Owen Cornec, Alfio Massimiliano Gliozzo

Abstract: We propose KnowGL, a tool that allows converting text into structured relational data represented as a set of ABox assertions compliant with the TBox of a given Knowledge Graph (KG), such as Wikidata. We address this problem as a sequence generation task by leveraging pre-trained sequence-to-sequence language models, e.g. BART. Given a sentence, we fine-tune such models to detect pairs of entity m… ▽ More We propose KnowGL, a tool that allows converting text into structured relational data represented as a set of ABox assertions compliant with the TBox of a given Knowledge Graph (KG), such as Wikidata. We address this problem as a sequence generation task by leveraging pre-trained sequence-to-sequence language models, e.g. BART. Given a sentence, we fine-tune such models to detect pairs of entity mentions and jointly generate a set of facts consisting of the full set of semantic annotations for a KG, such as entity labels, entity types, and their relationships. To showcase the capabilities of our tool, we build a web application consisting of a set of UI widgets that help users to navigate through the semantic data extracted from a given input text. We make the KnowGL model available at https://huggingface.co/ibm/knowgl-large. △ Less

Submitted 22 November, 2022; v1 submitted 25 October, 2022; originally announced October 2022.

Comments: AAAI-23 Demo Track

Prompt-guided Scene Generation for 3D Zero-Shot Learning

Authors: Majid Nasiri, Ali Cheraghian, Townim Faisal Chowdhury, Sahar Ahmadi, Morteza Saberi, Shafin Rahman

Abstract: Zero-shot learning on 3D point cloud data is a related underexplored problem compared to its 2D image counterpart. 3D data brings new challenges for ZSL due to the unavailability of robust pre-trained feature extraction models. To address this problem, we propose a prompt-guided 3D scene generation and supervision method that augments 3D data to learn the network better, exploring the complex inte… ▽ More Zero-shot learning on 3D point cloud data is a related underexplored problem compared to its 2D image counterpart. 3D data brings new challenges for ZSL due to the unavailability of robust pre-trained feature extraction models. To address this problem, we propose a prompt-guided 3D scene generation and supervision method that augments 3D data to learn the network better, exploring the complex interplay of seen and unseen objects. First, we merge point clouds of two 3D models in certain ways described by a prompt. The prompt acts like the annotation describing each 3D scene. Later, we perform contrastive learning to train our proposed architecture in an end-to-end manner. We argue that 3D scenes can relate objects more efficiently than single objects because popular language models (like BERT) can achieve high performance when objects appear in a context. Our proposed prompt-guided scene generation method encapsulates data augmentation and prompt-based annotation/captioning to improve 3D ZSL performance. We have achieved state-of-the-art ZSL and generalized ZSL performance on synthetic (ModelNet40, ModelNet10) and real-scanned (ScanOjbectNN) 3D object datasets. △ Less

Submitted 29 September, 2022; originally announced September 2022.

Re2G: Retrieve, Rerank, Generate

Authors: Michael Glass, Gaetano Rossiello, Md Faisal Mahbub Chowdhury, Ankita Rajaram Naik, Pengshan Cai, Alfio Gliozzo

Abstract: As demonstrated by GPT-3 and T5, transformers grow in capability as parameter spaces become larger and larger. However, for tasks that require a large amount of knowledge, non-parametric memory allows models to grow dramatically with a sub-linear increase in computational cost and GPU memory requirements. Recent models such as RAG and REALM have introduced retrieval into conditional generation. Th… ▽ More As demonstrated by GPT-3 and T5, transformers grow in capability as parameter spaces become larger and larger. However, for tasks that require a large amount of knowledge, non-parametric memory allows models to grow dramatically with a sub-linear increase in computational cost and GPU memory requirements. Recent models such as RAG and REALM have introduced retrieval into conditional generation. These models incorporate neural initial retrieval from a corpus of passages. We build on this line of research, proposing Re2G, which combines both neural initial retrieval and reranking into a BART-based sequence-to-sequence generation. Our reranking approach also permits merging retrieval results from sources with incomparable scores, enabling an ensemble of BM25 and neural initial retrieval. To train our system end-to-end, we introduce a novel variation of knowledge distillation to train the initial retrieval, reranker, and generation using only ground truth on the target sequence output. We find large gains in four diverse tasks: zero-shot slot filling, question answering, fact-checking, and dialog, with relative gains of 9% to 34% over the previous state-of-the-art on the KILT leaderboard. We make our code available as open source at https://github.com/IBM/kgi-slot-filling/tree/re2g. △ Less

Submitted 13 July, 2022; originally announced July 2022.

Comments: Accepted at NAACL 2022

Knowledge Graph Induction enabling Recommending and Trend Analysis: A Corporate Research Community Use Case

Authors: Nandana Mihindukulasooriya, Mike Sava, Gaetano Rossiello, Md Faisal Mahbub Chowdhury, Irene Yachbes, Aditya Gidh, Jillian Duckwitz, Kovit Nisar, Michael Santos, Alfio Gliozzo

Abstract: A research division plays an important role of driving innovation in an organization. Drawing insights, following trends, kee** abreast of new research, and formulating strategies are increasingly becoming more challenging for both researchers and executives as the amount of information grows in both velocity and volume. In this paper we present a use case of how a corporate research community,… ▽ More A research division plays an important role of driving innovation in an organization. Drawing insights, following trends, kee** abreast of new research, and formulating strategies are increasingly becoming more challenging for both researchers and executives as the amount of information grows in both velocity and volume. In this paper we present a use case of how a corporate research community, IBM Research, utilizes Semantic Web technologies to induce a unified Knowledge Graph from both structured and textual data obtained by integrating various applications used by the community related to research projects, academic papers, datasets, achievements and recognition. In order to make the Knowledge Graph more accessible to application developers, we identified a set of common patterns for exploiting the induced knowledge and exposed them as APIs. Those patterns were born out of user research which identified the most valuable use cases or user pain points to be alleviated. We outline two distinct scenarios: recommendation and analytics for business use. We will discuss these scenarios in detail and provide an empirical evaluation on entity recommendation specifically. The methodology used and the lessons learned from this work can be applied to other organizations facing similar challenges. △ Less

Submitted 15 September, 2022; v1 submitted 11 July, 2022; originally announced July 2022.

Comments: Accepted at ISWC 2022

MSC Class: 68T01; 68T30 ACM Class: I.2.7; I.2.4; H.5

A library of quantitative markers of seizure severity

Authors: Sarah J. Gascoigne, Leonard Waldmann, Mariella Panagiotopoulou, Fahmida Chowdhury, Alison Cronie, Beate Diehl, John S. Duncan, Jennifer Falconer, Yu Guan, Veronica Leach, Shona Livingstone, Christoforos Papasavvas, Ryan Faulder, Jess Blickwedel, Gabrielle M. Schroeder, Rhys H. Thomas, Kevin Wilson, Peter N. Taylor, Yujiang Wang

Abstract: Purpose: Understanding fluctuations of seizure severity within individuals is important for defining treatment outcomes and response to therapy, as well as develo** novel treatments for epilepsy. Current methods for grading seizure severity rely on qualitative interpretations from patients and clinicians. Quantitative measures of seizure severity would complement existing approaches, for EEG mon… ▽ More Purpose: Understanding fluctuations of seizure severity within individuals is important for defining treatment outcomes and response to therapy, as well as develo** novel treatments for epilepsy. Current methods for grading seizure severity rely on qualitative interpretations from patients and clinicians. Quantitative measures of seizure severity would complement existing approaches, for EEG monitoring, outcome monitoring, and seizure prediction. Therefore, we developed a library of quantitative electroencephalographic (EEG) markers that assess the spread and intensity of abnormal electrical activity during and after seizures. Methods: We analysed intracranial EEG (iEEG) recordings of 1056 seizures from 63 patients. For each seizure, we computed 16 markers of seizure severity that capture the signal magnitude, spread, duration, and post-ictal suppression of seizures. Results: Quantitative EEG markers of seizure severity distinguished focal vs. subclinical and focal vs. FTBTC seizures across patients. In individual patients, 71% had a moderate to large difference (ranksum r > 0.3) between focal and subclinical seizures in three or more markers. Circadian and longer-term changes in severity were found for 67% and 53% of patients, respectively. Conclusion: We demonstrate the feasibility of using quantitative iEEG markers to measure seizure severity. Our quantitative markers distinguish between seizure types and are therefore sensitive to established qualitative differences in seizure severity. Our results also suggest that seizure severity is modulated over different timescales. We envisage that our proposed seizure severity library will be expanded and updated in collaboration with the epilepsy research community to include more measures and modalities. △ Less

Submitted 27 January, 2023; v1 submitted 30 June, 2022; originally announced June 2022.

Comments: 30 pages main text, 18 pages supplementary. Six main figures, two supplementary figures

Driven spin dynamics enhances cryptochrome magnetoreception: Towards live quantum sensing

Authors: Luke D. Smith, Farhan T. Chowdhury, Iona Peasgood, Nahnsu Dawkins, Daniel R. Kattnig

Abstract: The mechanism underlying magnetoreception has long eluded explanation. A popular hypothesis attributes this sense to the quantum coherent spin dynamics of spin-selective recombination reactions of radical pairs in the protein cryptochrome. However, concerns about the validity of the hypothesis have been raised as unavoidable inter-radical interactions, such as strong electron-electron dipolar coup… ▽ More The mechanism underlying magnetoreception has long eluded explanation. A popular hypothesis attributes this sense to the quantum coherent spin dynamics of spin-selective recombination reactions of radical pairs in the protein cryptochrome. However, concerns about the validity of the hypothesis have been raised as unavoidable inter-radical interactions, such as strong electron-electron dipolar coupling, appear to suppress its sensitivity. We demonstrate that this can be overcome by driving the spin system through a modulation of the inter-radical distance. It is shown that this dynamical process markedly enhances geomagnetic field sensitivity in strongly coupled radical pairs via a Landau-Zener type transition between singlet and triplet states. These findings suggest that a "live" harmonically driven magnetoreceptor can be more sensitive than its "dead" static counterpart. △ Less

Submitted 15 June, 2022; originally announced June 2022.

Comments: 7 pages, 4 figures, in addition to Supporting Material of 15 pages and 12 figures

Journal ref: Journal of Physical Chemistry Letters 13, 10500 (2022)

Rethinking Task-Incremental Learning Baselines

Authors: Md Sazzad Hossain, Pritom Saha, Townim Faisal Chowdhury, Shafin Rahman, Fuad Rahman, Nabeel Mohammed

Abstract: It is common to have continuous streams of new data that need to be introduced in the system in real-world applications. The model needs to learn newly added capabilities (future tasks) while retaining the old knowledge (past tasks). Incremental learning has recently become increasingly appealing for this problem. Task-incremental learning is a kind of incremental learning where task identity of n… ▽ More It is common to have continuous streams of new data that need to be introduced in the system in real-world applications. The model needs to learn newly added capabilities (future tasks) while retaining the old knowledge (past tasks). Incremental learning has recently become increasingly appealing for this problem. Task-incremental learning is a kind of incremental learning where task identity of newly included task (a set of classes) remains known during inference. A common goal of task-incremental methods is to design a network that can operate on minimal size, maintaining decent performance. To manage the stability-plasticity dilemma, different methods utilize replay memory of past tasks, specialized hardware, regularization monitoring etc. However, these methods are still less memory efficient in terms of architecture growth or input data costs. In this study, we present a simple yet effective adjustment network (SAN) for task incremental learning that achieves near state-of-the-art performance while using minimal architectural size without using memory instances compared to previous state-of-the-art approaches. We investigate this approach on both 3D point cloud object (ModelNet40) and 2D image (CIFAR10, CIFAR100, MiniImageNet, MNIST, PermutedMNIST, notMNIST, SVHN, and FashionMNIST) recognition tasks and establish a strong baseline result for a fair comparison with existing methods. On both 2D and 3D domains, we also observe that SAN is primarily unaffected by different task orders in a task-incremental setting. △ Less

Submitted 23 May, 2022; originally announced May 2022.

Comments: Accepted in ICPR2022

Impact of reaction temperatures on the particle size of V2O5 synthesized by facile hydrothermal technique and their auspicious photocatalytic performance in dye degradation

Authors: M. A. Jalil, M. N. I. Khan, S. Mandal, F. -U. -Z. Chowdhury, M. M. Hossain, D. Jana, M. S. Alam, M. M. Uddin

Abstract: In this study, a complete study of the effect of hydrothermal reaction temperatures on the synthesis and physical properties of V2O5 using the green facile mild hydrothermal method has been performed with six different temperatures 100 °C to 200 °C, in the step of 20 °C. . The XRD pattern confirm the stable orthorhombic crystal structure of the synthesized samples at all reaction temperatures. The… ▽ More In this study, a complete study of the effect of hydrothermal reaction temperatures on the synthesis and physical properties of V2O5 using the green facile mild hydrothermal method has been performed with six different temperatures 100 °C to 200 °C, in the step of 20 °C. . The XRD pattern confirm the stable orthorhombic crystal structure of the synthesized samples at all reaction temperatures. The SEM and TEM images demonstrate the particle-like morphology, and these characterizations affirmed that the particles size became larger with the increase of reaction temperatures. The FTIR analysis is employed to study the functional groups, and the obtained results are consistent with the XRD analysis. The bandgap has been estimated at various reaction temperatures using UV-vis diffuse reflectance spectra (UV-DRS) and was found to be varied 2.09 eV to 2.15 eV that are suitable range to absorb a significant amount of visible light. The photocatalysis of methylene blue (MB) with synthesized samples has been accomplished to investigate photocatalytic efficiency. The pure V2O5 synthesized at lower reaction temperature (100 °C) possess a lower bandgap and, accordingly, higher photocatalytic efficiency. △ Less

Submitted 9 May, 2022; originally announced May 2022.

Comments: 19 pages

Intracranial EEG structure-function coupling predicts surgical outcomes in focal epilepsy

Authors: Nishant Sinha, John S. Duncan, Beate Diehl, Fahmida A. Chowdhury, Jane de Tisi, Anna Miserocchi, Andrew W. McEvoy, Kathryn A. Davis, Sjoerd B. Vos, Gavin P. Winston, Yujiang Wang, Peter N. Taylor

Abstract: Alterations to structural and functional brain networks have been reported across many neurological conditions. However, the relationship between structure and function -- their coupling -- is relatively unexplored, particularly in the context of an intervention. Epilepsy surgery alters the brain structure and networks to control the functional abnormality of seizures. Given that surgery is a stru… ▽ More Alterations to structural and functional brain networks have been reported across many neurological conditions. However, the relationship between structure and function -- their coupling -- is relatively unexplored, particularly in the context of an intervention. Epilepsy surgery alters the brain structure and networks to control the functional abnormality of seizures. Given that surgery is a structural modification aiming to alter the function, we hypothesized that stronger structure-function coupling preoperatively is associated with a greater chance of post-operative seizure control. We constructed structural and functional brain networks in 39 subjects with medication-resistant focal epilepsy using data from intracranial EEG (pre-surgery), structural MRI (pre-and post-surgery), and diffusion MRI (pre-surgery). We investigated pre-operative structure-function coupling at two spatial scales a) at the global iEEG network level and b) at the resolution of individual iEEG electrode contacts using virtual surgeries. At global network level, seizure-free individuals had stronger structure-function coupling pre-operatively than those that were not seizure-free regardless of the choice of interictal segment or frequency band. At the resolution of individual iEEG contacts, the virtual surgery approach provided complementary information to localize epileptogenic tissues. In predicting seizure outcomes, structure-function coupling measures were more important than clinical attributes, and together they predicted seizure outcomes with an accuracy of 85% and sensitivity of 87%. The underlying assumption that the structural changes induced by surgery translate to the functional level to control seizures is valid when the structure-functional coupling is strong. Map** the regions that contribute to structure-functional coupling using virtual surgeries may help aid surgical planning. △ Less

Submitted 17 April, 2022; originally announced April 2022.

KGI: An Integrated Framework for Knowledge Intensive Language Tasks

Authors: Md Faisal Mahbub Chowdhury, Michael Glass, Gaetano Rossiello, Alfio Gliozzo, Nandana Mihindukulasooriya

Abstract: In this paper, we present a system to showcase the capabilities of the latest state-of-the-art retrieval augmented generation models trained on knowledge-intensive language tasks, such as slot filling, open domain question answering, dialogue, and fact-checking. Moreover, given a user query, we show how the output from these different models can be combined to cross-examine the outputs of each oth… ▽ More In this paper, we present a system to showcase the capabilities of the latest state-of-the-art retrieval augmented generation models trained on knowledge-intensive language tasks, such as slot filling, open domain question answering, dialogue, and fact-checking. Moreover, given a user query, we show how the output from these different models can be combined to cross-examine the outputs of each other. Particularly, we show how accuracy in dialogue can be improved using the question answering model. We are also releasing all models used in the demo as a contribution of this paper. A short video demonstrating the system is available at https://ibm.box.com/v/emnlp2022-demo. △ Less

Submitted 21 September, 2022; v1 submitted 8 April, 2022; originally announced April 2022.

Comments: EMNLP 2022 Demo Track

Quantum Control of Spin Qubits Using Nanomagnets

Authors: Mohamad Niknam, Md Fahim F. Chowdhury, Md Mahadi Rajib, Walid Al Misba, Robert N. Schwartz, Kang L. Wang, Jayasimha Atulasimha, Louis-S. Bouchard

Abstract: Single-qubit gates are essential components of a universal quantum computer. Without selective addressing of individual qubits, scalable implementation of quantum algorithms is not possible. When the qubits are discrete points or regions on a lattice, the selective addressing of magnetic spin qubits at the nanoscale remains a challenge due to the difficulty of localizing and confining a classical… ▽ More Single-qubit gates are essential components of a universal quantum computer. Without selective addressing of individual qubits, scalable implementation of quantum algorithms is not possible. When the qubits are discrete points or regions on a lattice, the selective addressing of magnetic spin qubits at the nanoscale remains a challenge due to the difficulty of localizing and confining a classical divergence-free field to a small volume of space. Herein we propose a new technique for addressing spin qubits using voltage-control of nanoscale magnetism, exemplified by the use of voltage control of magnetic anisotropy (VCMA). We show that by tuning the frequency of the nanomagnet's electric field drive to the Larmor frequency of the spins confined to a nanoscale volume, and by modulating the phase of the drive, single-qubit quantum gates with fidelities approaching those for fault-tolerant quantum computing can be implemented. Such single-qubit gate operations have the advantage of remarkable energy efficiency, requiring only tens of femto-Joules per gate operation, and lossless, purely magnetic field control (no E-field over the target volume). Their physical realization is also straightforward using existing foundry manufacturing techniques. △ Less

Submitted 30 March, 2022; originally announced March 2022.

Comments: 10 pages, 6 figures

Applying a Generic Sequence-to-Sequence Model for Simple and Effective Keyphrase Generation

Authors: Md Faisal Mahbub Chowdhury, Gaetano Rossiello, Michael Glass, Nandana Mihindukulasooriya, Alfio Gliozzo

Abstract: In recent years, a number of keyphrase generation (KPG) approaches were proposed consisting of complex model architectures, dedicated training paradigms and decoding strategies. In this work, we opt for simplicity and show how a commonly used seq2seq language model, BART, can be easily adapted to generate keyphrases from the text in a single batch computation using a simple training procedure. Emp… ▽ More In recent years, a number of keyphrase generation (KPG) approaches were proposed consisting of complex model architectures, dedicated training paradigms and decoding strategies. In this work, we opt for simplicity and show how a commonly used seq2seq language model, BART, can be easily adapted to generate keyphrases from the text in a single batch computation using a simple training procedure. Empirical results on five benchmarks show that our approach is as good as the existing state-of-the-art KPG systems, but using a much simpler and easy to deploy framework. △ Less

Submitted 13 January, 2022; originally announced January 2022.

Skyrmion based energy efficient straintronic physical reservoir computing

Authors: Md Mahadi Rajib, Walid Al Misba, Md. Fahim F. Chowdhury, Muhammad Sabbir Alam, Jayasimha Atulasimha

Abstract: Physical Reservoir Computing (PRC) is an unconventional computing paradigm, which exploits nonlinear dynamics of reservoir blocks to perform recognition and classification tasks. Here we show with simulations that patterned thin films hosting several skyrmions, particularly one, two, four and nine skyrmions, can implement energy efficient reservoir computing. This reservoir computing (RC) block is… ▽ More Physical Reservoir Computing (PRC) is an unconventional computing paradigm, which exploits nonlinear dynamics of reservoir blocks to perform recognition and classification tasks. Here we show with simulations that patterned thin films hosting several skyrmions, particularly one, two, four and nine skyrmions, can implement energy efficient reservoir computing. This reservoir computing (RC) block is based on nonlinear breathing dynamics of skyrmions, which are coupled to each other through dipole interaction and spin waves, in response to a voltage generated strain. This nonlinear and coupled magnetization dynamics is exploited to perform temporal pattern recognition. Two performance metrics, namely Short-Term Memory (STM) and Parity Check (PC) capacity are studied to demonstrate the potential of such skyrmion based PRC in addition to showing it can classify sine and square waves with 100% accuracy. Furthermore, our study demonstrates that nonlinear magnetization dynamics and interaction through spin waves and dipole coupling have a strong influence on STM and PC capacity, thus explaining the role of physical interactions in a dynamical system on its ability to perform Reservoir Computing (RC). △ Less

Submitted 13 April, 2022; v1 submitted 27 December, 2021; originally announced December 2021.

Focused Surface Acoustic Wave induced nano-oscillator based reservoir computing

Authors: Md. Fahim F. Chowdhury, Walid Al Misba, Md Mahadi Rajib, Alexander J. Edwards, Dhritiman Bhattacharya, Joseph S. Friedman, Jayasimha Atulasimha

Abstract: We demonstrate using micromagnetic simulations that a nanomagnet array excited by Surface Acoustic Waves (SAWs) can work as a reservoir that can classify sine and square waves with high accuracy. To evaluate memory effect and computing capability, we study the Short-Term Memory (STM) and Parity Check (PC) capacities respectively. The simulated nanomagnet array has an input nanomagnet that is excit… ▽ More We demonstrate using micromagnetic simulations that a nanomagnet array excited by Surface Acoustic Waves (SAWs) can work as a reservoir that can classify sine and square waves with high accuracy. To evaluate memory effect and computing capability, we study the Short-Term Memory (STM) and Parity Check (PC) capacities respectively. The simulated nanomagnet array has an input nanomagnet that is excited with focused SAW and coupled to several nanomagnets, seven of which serve as output nanomagnets. The SAW has a carrier frequency of 4 GHz whose amplitude is modulated to provide different inputs of sine and square waves of 100 MHz frequency. The responses of the selected output nanomagnets are processed by reading the envelope of their magnetization state, which is used to train the output weights using regression method (e.g. Moore-Penrose pseudoinverse operation). For classification, a random sequence of 100 square and sine wave samples are used, of which 80 % are used for training, and the rest of the samples used for testing. We achieve 100 % training accuracy and 100 % testing accuracy for different combination of nanomagnets as outputs. Further, the STM and PC is calculated to be ~ 5.5 bits and ~ 5.3 bits respectively, which is indicative of the proposed acoustically driven nanomagnet oscillator array being well suited for physical reservoir computing applications. Finally, the ability to use high frequency (4GHz, wavelength ~1 micron) SAW makes the device scalable to small dimensions, while the ability to modulate the envelope at lower frequency (100 MHz) adds flexibility to encode different signals beyond the sine and square waves demonstrated here. △ Less

Submitted 9 April, 2022; v1 submitted 7 December, 2021; originally announced December 2021.

A Systematic Literature Review on Wearable Health Data Publishing under Differential Privacy

Authors: Munshi Saifuzzaman, Tajkia Nuri Ananna, Mohammad Jabed Morshed Chowdhury, Md Sadek Ferdous, Farida Chowdhury

Abstract: Wearable devices generate different types of physiological data about the individuals. These data can provide valuable insights for medical researchers and clinicians that cannot be availed through traditional measures. Researchers have historically relied on survey responses or observed behavior. Interestingly, physiological data can provide a richer amount of user cognition than that obtained fr… ▽ More Wearable devices generate different types of physiological data about the individuals. These data can provide valuable insights for medical researchers and clinicians that cannot be availed through traditional measures. Researchers have historically relied on survey responses or observed behavior. Interestingly, physiological data can provide a richer amount of user cognition than that obtained from any other sources, including the user himself. Therefore, the inexpensive consumer-grade wearable devices have become a point of interest for the health researchers. In addition, they are also used in continuous remote health monitoring and sometimes by the insurance companies. However, the biggest concern for such kind of use cases is the privacy of the individuals. There are a few privacy mechanisms, such as abstraction and k-anonymity, are widely used in information systems. Recently, Differential Privacy (DP) has emerged as a proficient technique to publish privacy sensitive data, including data from wearable devices. In this paper, we have conducted a Systematic Literature Review (SLR) to identify, select and critically appraise researches in DP as well as to understand different techniques and exiting use of DP in wearable data publishing. Based on our study we have identified the limitations of proposed solutions and provided future directions. △ Less

Submitted 15 September, 2021; originally announced September 2021.

Comments: 25 pages

Seizure pathways and seizure durations can vary independently within individual patients with focal epilepsy

Authors: Gabrielle M. Schroeder, Fahmida A. Chowdhury, Mark J. Cook, Beate Diehl, John S. Duncan, Philippa J. Karoly, Peter N. Taylor, Yujiang Wang

Abstract: A seizure's electrographic dynamics are characterised by its spatiotemporal evolution, also termed dynamical "pathway" and the time it takes to complete that pathway, which results in the seizure's duration. Both seizure pathways and durations can vary within the same patient, producing seizures with different dynamics, severity, and clinical implications. However, it is unclear whether seizures f… ▽ More A seizure's electrographic dynamics are characterised by its spatiotemporal evolution, also termed dynamical "pathway" and the time it takes to complete that pathway, which results in the seizure's duration. Both seizure pathways and durations can vary within the same patient, producing seizures with different dynamics, severity, and clinical implications. However, it is unclear whether seizures following the same pathway will have the same duration or if these features can vary independently. We compared within-subject variability in these seizure features using 1) epilepsy monitoring unit intracranial EEG (iEEG) recordings of 31 patients (mean 6.7 days, 16.5 seizures/subject), 2) NeuroVista chronic iEEG recordings of 10 patients (mean 521.2 days, 252.6 seizures/subject), and 3) chronic iEEG recordings of 3 dogs with focal-onset seizures (mean 324.4 days, 62.3 seizures/subject). While the strength of the relationship between seizure pathways and durations was highly subject-specific, in most subjects, changes in seizure pathways were only weakly to moderately associated with differences in seizure durations. The relationship between seizure pathways and durations was weakened by seizures that 1) had a common pathway, but different durations ("elastic pathways"), or 2) had similar durations, but followed different pathways ("duplicate durations"). Even in subjects with distinct populations of short and long seizures, seizure durations were not a reliable indicator of different seizure pathways. These findings suggest that seizure pathways and durations are modulated by different processes. Uncovering such modulators may reveal novel therapeutic targets for reducing seizure duration and severity. △ Less

Submitted 14 September, 2021; originally announced September 2021.

Robust Retrieval Augmented Generation for Zero-shot Slot Filling

Authors: Michael Glass, Gaetano Rossiello, Md Faisal Mahbub Chowdhury, Alfio Gliozzo

Abstract: Automatically inducing high quality knowledge graphs from a given collection of documents still remains a challenging problem in AI. One way to make headway for this problem is through advancements in a related task known as slot filling. In this task, given an entity query in form of [Entity, Slot, ?], a system is asked to fill the slot by generating or extracting the missing value exploiting evi… ▽ More Automatically inducing high quality knowledge graphs from a given collection of documents still remains a challenging problem in AI. One way to make headway for this problem is through advancements in a related task known as slot filling. In this task, given an entity query in form of [Entity, Slot, ?], a system is asked to fill the slot by generating or extracting the missing value exploiting evidence extracted from relevant passage(s) in the given document collection. The recent works in the field try to solve this task in an end-to-end fashion using retrieval-based language models. In this paper, we present a novel approach to zero-shot slot filling that extends dense passage retrieval with hard negatives and robust training procedures for retrieval augmented generation models. Our model reports large improvements on both T-REx and zsRE slot filling datasets, improving both passage retrieval and slot value generation, and ranking at the top-1 position in the KILT leaderboard. Moreover, we demonstrate the robustness of our system showing its domain adaptation capability on a new variant of the TACRED dataset for slot filling, through a combination of zero/few-shot learning. We release the source code and pre-trained models. △ Less

Submitted 13 September, 2021; v1 submitted 31 August, 2021; originally announced August 2021.

Comments: Accepted at EMNLP 2021. arXiv admin note: substantial text overlap with arXiv:2104.08610

Normative brain map** of interictal intracranial EEG to localise epileptogenic tissue

Authors: Peter N Taylor, Christoforos A Papasavvas, Thomas W Owen, Gabrielle M Schroeder, Frances E Hutchings, Fahmida A Chowdhury, Beate Diehl, John S Duncan, Andrew W McEvoy, Anna Miserocchi, Jane de Tisi, Sjoerd B Vos, Matthew C Walker, Yujiang Wang

Abstract: The identification of abnormal electrographic activity is important in a wide range of neurological disorders, including epilepsy for localising epileptogenic tissue. However, this identification may be challenging during non-seizure (interictal) periods, especially if abnormalities are subtle compared to the repertoire of possible healthy brain dynamics. Here, we investigate if such interictal ab… ▽ More The identification of abnormal electrographic activity is important in a wide range of neurological disorders, including epilepsy for localising epileptogenic tissue. However, this identification may be challenging during non-seizure (interictal) periods, especially if abnormalities are subtle compared to the repertoire of possible healthy brain dynamics. Here, we investigate if such interictal abnormalities become more salient by quantitatively accounting for the range of healthy brain dynamics in a location-specific manner. To this end, we constructed a normative map of brain dynamics, in terms of relative band power, from interictal intracranial recordings from 234 subjects (21,598 electrode contacts). We then compared interictal recordings from 62 patients with epilepsy to the normative map to identify abnormal regions. We hypothesised that if the most abnormal regions were spared by surgery, then patients would be more likely to experience continued seizures post-operatively. We first confirmed that the spatial variations of band power in the normative map across brain regions were consistent with healthy variations reported in the literature. Second, when accounting for the normative variations, regions which were spared by surgery were more abnormal than those resected only in patients with persistent post-operative seizures (t=-3.6, p=0.0003), confirming our hypothesis. Third, we found that this effect discriminated patient outcomes (AUC=0.75 p=0.0003). Normative map** is a well-established practice in neuroscientific research. Our study suggests that this approach is feasible to detect interictal abnormalities in intracranial EEG, and of potential clinical value to identify pathological tissue in epilepsy. Finally, we make our normative intracranial map publicly available to facilitate future investigations in epilepsy and beyond. △ Less

Submitted 7 January, 2022; v1 submitted 10 May, 2021; originally announced May 2021.

Comments: 25 pages, 6 figures

Impact of particle size on the magnetic properties of highly crystalline Yb3+ substituted Ni-Zn nanoferrites

Authors: N. Jahan, M. M. Uddin, M. N. I. Khan, F. -U. -Z. Chowdhury, M. R. Hasan, H. N. Das, M. M. Hossain

Abstract: Yb-substituted Ni-Zn ferrites have been synthesized using sol-gel auto combustion method. The structural characterization of the compositions has been performed by X-ray diffraction analysis, field emission scanning electron microscopy (FESEM), quantum design physical properties measurement system (PPMS). That ensured the formation of single phase cubic spinel structure. Crystallite and average gr… ▽ More Yb-substituted Ni-Zn ferrites have been synthesized using sol-gel auto combustion method. The structural characterization of the compositions has been performed by X-ray diffraction analysis, field emission scanning electron microscopy (FESEM), quantum design physical properties measurement system (PPMS). That ensured the formation of single phase cubic spinel structure. Crystallite and average grain size are calculated and found to decrease with increasing Yb3+ contents. Saturation magnetization and Bohr magnetic moment decrease while the coercivity increases with the increase in Yb3+ contents successfully explained by the Neels collinear two sub-lattice model and critical size effect, respectively. Critical particle size has been estimated at 6.4 nm, the transition point between single domain regime (below the critical size) and multi-domain regime (beyond the critical size). Curie temperature reduces due to the weakening of A-O-B super exchange interaction and redistribution of cations, confirmed by the M-T graph. The compositions retain ferromagnetic ordered structured below Curie temperature and above Curie temperature, it becomes paramagnetic, making them plausible candidates for high temperature magnetic device applications. The relative quality factor peak is obtained at a very high frequency, indicating the compositions could also be applicable for high frequency magnetic device applications. △ Less

Submitted 17 April, 2021; originally announced April 2021.

Zero-Shot Learning on 3D Point Cloud Objects and Beyond

Authors: Ali Cheraghian, Shafinn Rahman, Townim F. Chowdhury, Dylan Campbell, Lars Petersson

Abstract: Zero-shot learning, the task of learning to recognize new classes not seen during training, has received considerable attention in the case of 2D image classification. However, despite the increasing ubiquity of 3D sensors, the corresponding 3D point cloud classification problem has not been meaningfully explored and introduces new challenges. In this paper, we identify some of the challenges and… ▽ More Zero-shot learning, the task of learning to recognize new classes not seen during training, has received considerable attention in the case of 2D image classification. However, despite the increasing ubiquity of 3D sensors, the corresponding 3D point cloud classification problem has not been meaningfully explored and introduces new challenges. In this paper, we identify some of the challenges and apply 2D Zero-Shot Learning (ZSL) methods in the 3D domain to analyze the performance of existing models. Then, we propose a novel approach to address the issues specific to 3D ZSL. We first present an inductive ZSL process and then extend it to the transductive ZSL and Generalized ZSL (GZSL) settings for 3D point cloud classification. To this end, a novel loss function is developed that simultaneously aligns seen semantics with point cloud features and takes advantage of unlabeled test data to address some known issues (e.g., the problems of domain adaptation, hubness, and data bias). While designed for the particularities of 3D point cloud classification, the method is shown to also be applicable to the more common use-case of 2D image classification. An extensive set of experiments is carried out, establishing state-of-the-art for ZSL and GZSL on synthetic (ModelNet40, ModelNet10, McGill) and real (ScanObjectNN) 3D point cloud datasets. △ Less

Submitted 11 April, 2021; originally announced April 2021.

Comments: arXiv admin note: substantial text overlap with arXiv:1912.07161

Passive frustrated nanomagnet reservoir computing

Authors: Alexander J. Edwards, Dhritiman Bhattacharya, Peng Zhou, Nathan R. McDonald, Walid Al Misba, Lisa Loomis, Felipe Garcia-Sanchez, Naimul Hassan, Xuan Hu, Md. Fahim Chowdhury, Clare D. Thiem, Jayasimha Atulasimha, Joseph S. Friedman

Abstract: Reservoir computing (RC) has received recent interest because reservoir weights do not need to be trained, enabling extremely low-resource consumption implementations, which could have a transformative impact on edge computing and in-situ learning where resources are severely constrained. Ideally, a natural hardware reservoir should be passive, minimal, expressive, and feasible; to date, proposed… ▽ More Reservoir computing (RC) has received recent interest because reservoir weights do not need to be trained, enabling extremely low-resource consumption implementations, which could have a transformative impact on edge computing and in-situ learning where resources are severely constrained. Ideally, a natural hardware reservoir should be passive, minimal, expressive, and feasible; to date, proposed hardware reservoirs have had difficulty meeting all of these criteria. We therefore propose a reservoir that meets all of these criteria by leveraging the passive interactions of dipole-coupled, frustrated nanomagnets. The frustration significantly increases the number of stable reservoir states, enriching reservoir dynamics, and as such these frustrated nanomagnets fulfill all of the criteria for a natural hardware reservoir. We likewise propose a complete frustrated nanomagnet reservoir computing (NMRC) system with low-power complementary metal-oxide semiconductor (CMOS) circuitry to interface with the reservoir, and initial experimental results demonstrate the reservoir's feasibility. The reservoir is verified with micromagnetic simulations on three separate tasks demonstrating expressivity. The proposed system is compared with a CMOS echo-state-network (ESN), demonstrating an overall resource decrease by a factor of over 10,000,000, demonstrating that because NMRC is naturally passive and minimal it has the potential to be extremely resource efficient. △ Less

Submitted 16 September, 2022; v1 submitted 16 March, 2021; originally announced March 2021.

Micrometer-thick, atomically random Si0.06Ge0.90Sn0.04 for silicon-integrated infrared optoelectronics

Authors: Simone Assali, Anis Attiaoui, Sebastian Koelling, Mahmoud R. M. Atalla, Aashish Kumar, Jérôme Nicolas, Faqrul A. Chowdhury, Cédric Lemieux-Leduc, Oussama Moutanabbir

Abstract: A true monolithic infrared photonics platform is within reach if strain and bandgap energy can be independently engineered in SiGeSn semiconductors. Herein, we investigate the structural and optoelectronic properties of a 1.5 μm-thick Si0.06Ge0.90Sn0.04 layer that is nearly lattice-matched to a Ge on Si substrate. Atomic-level studies demonstrate high crystalline quality and uniform composition an… ▽ More A true monolithic infrared photonics platform is within reach if strain and bandgap energy can be independently engineered in SiGeSn semiconductors. Herein, we investigate the structural and optoelectronic properties of a 1.5 μm-thick Si0.06Ge0.90Sn0.04 layer that is nearly lattice-matched to a Ge on Si substrate. Atomic-level studies demonstrate high crystalline quality and uniform composition and show no sign of short-range ordering and clusters. Room temperature spectroscopic ellipsometry and transmission measurements show direct bandgap absorption at 0.83 eV and a reduced indirect bandgap absorption at lower energies. Si0.06Ge0.90Sn0.04 photoconductive devices operating at room temperature exhibit dark current and spectral responsivity (1 A/W below 1.5 μm wavelengths) similar to Ge on Si devices, with the advantage of a near-infrared band gap tunable by alloy composition. These results underline the relevance of SiGeSn semiconductors in implementing a group IV material platform for silicon-integrated infrared optoelectronics. △ Less

Submitted 12 August, 2022; v1 submitted 3 March, 2021; originally announced March 2021.

Examining Factors Associated with Twitter Account Suspension Following the 2020 U.S. Presidential Election

Authors: Farhan Asif Chowdhury, Dheeman Saha, Md Rashidul Hasan, Koustuv Saha, Abdullah Mueen

Abstract: Online social media enables mass-level, transparent, and democratized discussion on numerous socio-political issues. Due to such openness, these platforms often endure manipulation and misinformation - leading to negative impacts. To prevent such harmful activities, platform moderators employ countermeasures to safeguard against actors violating their rules. However, the correlation between public… ▽ More Online social media enables mass-level, transparent, and democratized discussion on numerous socio-political issues. Due to such openness, these platforms often endure manipulation and misinformation - leading to negative impacts. To prevent such harmful activities, platform moderators employ countermeasures to safeguard against actors violating their rules. However, the correlation between publicly outlined policies and employed action is less clear to general people. In this work, we examine violations and subsequent moderation related to the 2020 U.S. President Election discussion on Twitter, a popular micro-blogging site. We focus on quantifying plausible reasons for the suspension, drawing on Twitter's rules and policies by identifying suspended users (Case) and comparing their activities and properties with (yet) non-suspended (Control) users. Using a dataset of 240M election-related tweets made by 21M unique users, we observe that Suspended users violate Twitter's rules at a higher rate (statistically significant) than Control users across all the considered aspects - hate speech, offensiveness, spamming, and civic integrity. Moreover, through the lens of Twitter's suspension mechanism, we qualitatively examine the targeted topics for manipulation. △ Less

Submitted 23 January, 2021; originally announced January 2021.

A Study of Password Security Factors among Bangladeshi Government Websites

Authors: Adil Ahmed Chowdhury, Farida Chowdhury, Md Sadek Ferdous

Abstract: The Government of Bangladesh is aggressively transforming its public service landscape by transforming public services into online services via a number of websites. The motivation is that this would be a catalyst for a transformative change in every aspect of citizen life. Some web services must be protected from any unauthorised usages and passwords remain the most widely used credential mechani… ▽ More The Government of Bangladesh is aggressively transforming its public service landscape by transforming public services into online services via a number of websites. The motivation is that this would be a catalyst for a transformative change in every aspect of citizen life. Some web services must be protected from any unauthorised usages and passwords remain the most widely used credential mechanism for this purpose. However, if passwords are not adopted properly, they can be a cause for security breach. That is why it is important to study different aspects of password security on different websites. In this paper, we present a study of password security among 36 different Bangladeshi government websites against six carefully chosen password security heuristics. This study is the first of its kind in this domain and offers interesting insights. For example, many websites have not adopted proper security measures with respect to security. There is no password construction guideline adopted by many websites, thus creating a barrier for users to select a strong password. Some of them allow supposedly weak passwords and still do not utilise a secure HTTPS channel to transmit information over the Internet. △ Less

Submitted 3 December, 2020; originally announced December 2020.

Comments: Accepted for publication in the 23rd International Conference on Computer and Information Technology (ICCIT), 19-21 December, 2020

Template Controllable keywords-to-text Generation

Authors: Abhijit Mishra, Md Faisal Mahbub Chowdhury, Sagar Manohar, Dan Gutfreund, Karthik Sankaranarayanan

Abstract: This paper proposes a novel neural model for the understudied task of generating text from keywords. The model takes as input a set of un-ordered keywords, and part-of-speech (POS) based template instructions. This makes it ideal for surface realization in any NLG setup. The framework is based on the encode-attend-decode paradigm, where keywords and templates are encoded first, and the decoder jud… ▽ More This paper proposes a novel neural model for the understudied task of generating text from keywords. The model takes as input a set of un-ordered keywords, and part-of-speech (POS) based template instructions. This makes it ideal for surface realization in any NLG setup. The framework is based on the encode-attend-decode paradigm, where keywords and templates are encoded first, and the decoder judiciously attends over the contexts derived from the encoded keywords and templates to generate the sentences. Training exploits weak supervision, as the model trains on a large amount of labeled data with keywords and POS based templates prepared through completely automatic means. Qualitative and quantitative performance analyses on publicly available test-data in various domains reveal our system's superiority over baselines, built using state-of-the-art neural machine translation and controllable transfer techniques. Our approach is indifferent to the order of input keywords. △ Less

Submitted 7 November, 2020; originally announced November 2020.

Influence of Yb3+on the structural, electrical and optical properties of sol-gel synthesized Ni-Zn nanoferrites

Authors: N. Jahan, M. N. I. Khan, F. -U. -Z. Chowdhury, A. K. M. Akhter Hossain, S. M. Hoque, M. A. Matin, M. N. Hossain, M. M. Hossain, M. M. Uddin

Abstract: Polycrystalline Yb substituted NiZn nanoferrites with the compositions of Ni0.5Zn0.5YbxFe2-xO4 (x= 0.00, 0.04, 0.08, 0.12, 0.16 and 0.20) have been synthesized using sol gel auto combustion technique. Single phase cubic spinel structure has been confirmed by the X ray diffraction (XRD) patterns. Larger lattice constants of the compositions are found with increasing Yb3+ concentration while the ave… ▽ More Polycrystalline Yb substituted NiZn nanoferrites with the compositions of Ni0.5Zn0.5YbxFe2-xO4 (x= 0.00, 0.04, 0.08, 0.12, 0.16 and 0.20) have been synthesized using sol gel auto combustion technique. Single phase cubic spinel structure has been confirmed by the X ray diffraction (XRD) patterns. Larger lattice constants of the compositions are found with increasing Yb3+ concentration while the average grain size (52 to 18 nm) has noticeable decrease as Yb3+ content is increased. The presence of all existing elements as well as the purity of the samples has also been confirmed from energy dispersive X ray spectroscopic (EDS) analysis. Frequency dependent dielectric constant, dielectric loss, dielectric relaxation time, AC and DC resistivity of the compositions have also been examined at room temperature. The DC resistivity value is found in the order of 10 to power 10 (omega-cm) which is at least four orders greater than the ferrites prepared by conventional method. This larger value of resistivity attributes due to very small grain size and successfully explained using the Verwey and deBoer hop** conduction model. The contribution of grain and grain boundary resistance has been elucidated using Cole Cole plot. The study of temperature dependent DC resistivity confirms the semiconducting nature of all titled compositions wherein bandgap (optical) increases from 2.73 eV to 3.25 eV with the increase of Yb content. The high value of resistivity is of notable achievement for the compositions that make them a potential candidate for implication in the high frequency applications where reduction of eddy current loss is highly required. △ Less

Submitted 25 June, 2020; originally announced June 2020.

Comments: 27 pages

O(1) Communication for Distributed SGD through Two-Level Gradient Averaging

Authors: Subhadeep Bhattacharya, Weikuan Yu, Fahim Tahmid Chowdhury

Abstract: Large neural network models present a hefty communication challenge to distributed Stochastic Gradient Descent (SGD), with a communication complexity of O(n) per worker for a model of n parameters. Many sparsification and quantization techniques have been proposed to compress the gradients, some reducing the communication complexity to O(k), where k << n. In this paper, we introduce a strategy cal… ▽ More Large neural network models present a hefty communication challenge to distributed Stochastic Gradient Descent (SGD), with a communication complexity of O(n) per worker for a model of n parameters. Many sparsification and quantization techniques have been proposed to compress the gradients, some reducing the communication complexity to O(k), where k << n. In this paper, we introduce a strategy called two-level gradient averaging (A2SGD) to consolidate all gradients down to merely two local averages per worker before the computation of two global averages for an updated model. A2SGD also retains local errors to maintain the variance for fast convergence. Our theoretical analysis shows that A2SGD converges similarly like the default distributed SGD algorithm. Our evaluation validates the theoretical conclusion and demonstrates that A2SGD significantly reduces the communication traffic per worker, and improves the overall training time of LSTM-PTB by 3.2x and 23.2x, respectively, compared to Top-K and QSGD. To the best of our knowledge, A2SGD is the first to achieve O(1) communication complexity per worker for distributed SGD. △ Less

Submitted 15 June, 2020; v1 submitted 12 June, 2020; originally announced June 2020.

COVID-19: Social Media Sentiment Analysis on Reopening

Authors: Mohammed Emtiaz Ahmed, Md Rafiqul Islam Rabin, Farah Naz Chowdhury

Abstract: The novel coronavirus (COVID-19) pandemic is the most talked topic in social media platforms in 2020. People are using social media such as Twitter to express their opinion and share information on a number of issues related to the COVID-19 in this stay at home order. In this paper, we investigate the sentiment and emotion of peoples in the United States on the subject of reopening. We choose the… ▽ More The novel coronavirus (COVID-19) pandemic is the most talked topic in social media platforms in 2020. People are using social media such as Twitter to express their opinion and share information on a number of issues related to the COVID-19 in this stay at home order. In this paper, we investigate the sentiment and emotion of peoples in the United States on the subject of reopening. We choose the social media platform Twitter for our analysis and study the Tweets to discover the sentimental perspective, emotional perspective, and triggering words towards the reopening. During this COVID-19 pandemic, researchers have made some analysis on various social media dataset regarding lockdown and stay at home. However, in our analysis, we are particularly interested to analyse public sentiment on reopening. Our major finding is that when all states resorted to lockdown in March, people showed dominant emotion of fear, but as reopening starts people have less fear. While this may be true, due to this reopening phase daily positive cases are rising compared to the lockdown situation. Overall, people have a less negative sentiment towards the situation of reopening. △ Less

Submitted 1 June, 2020; originally announced June 2020.

Comments: 8 pages, 4 figures, 1 table

Mechanical behavior, enhanced dc resistivity, energy band gap and high temperature magnetic properties of Y-substituted Mg-Zn ferrites

Authors: M. A. Ali, M. N. I. Khan, M. M. Hossain, F. -U. -Z. Chowdhury, M. N. Hossain, R. Rashid, M. A. Hakim, S. M. Hoque, M. M. Uddin

Abstract: We report the synthesis of Y-substituted Mg-Zn ferrites using conventional standard ceramic technique. XRD patterns confirm the single phase cubic spinel structure up to x = 0.03 and appearance of a secondary phase of YFeO3for higher Y contents. FESEM images depict the distribution of grains and EDS spectra confirmed the absence of any unwanted element. Completion of solid state reaction and forma… ▽ More We report the synthesis of Y-substituted Mg-Zn ferrites using conventional standard ceramic technique. XRD patterns confirm the single phase cubic spinel structure up to x = 0.03 and appearance of a secondary phase of YFeO3for higher Y contents. FESEM images depict the distribution of grains and EDS spectra confirmed the absence of any unwanted element. Completion of solid state reaction and formation of spinel structure has been revealed from FTIR spectra. The FTIR data along with lattice constant, bulk density and porosity were further used to calculate the stiffness constant (Cij), elastic constant and Debye temperatures. Mechanical stability of all studied compositions is confirmed from Cij using Born stability conditions. Brittleness and isotropic nature are also confirmed using Poisson ratio and anisotropy constants, respectively. The enhancement of dc electrical resistivity with Y content is observed. The energy band gap (increased with Y contents) is found in good agreement with dc electrical resistivity. Ferrimagnetic to paramagnetic phase change has been observed from the field dependent high temperature magnetization curves. The magnetic moments and saturation magnetization were found to be decreased with increasing temperature. The Curie temperature (Tc) has been measured from temperature dependent magnetic moment (M-T) and initial permeability and found to be in good agreement with each other. Decrease in Tc with Y content is due to redistribution of cations and weakening of the exchange coupling constant. The magnetic phase transition has been analyzed by Arrott plot and found to have second order phase transition. The dc resistivity endorses the prepared ferrites are suitable for high frequency and high temperature magnetic device applications as well. △ Less

Submitted 20 January, 2020; originally announced January 2020.

Comments: 24 pages, Submitted to Materials Research Express

Hypernym Detection Using Strict Partial Order Networks

Authors: Sarthak Dash, Md Faisal Mahbub Chowdhury, Alfio Gliozzo, Nandana Mihindukulasooriya, Nicolas Rodolfo Fauceglia

Abstract: This paper introduces Strict Partial Order Networks (SPON), a novel neural network architecture designed to enforce asymmetry and transitive properties as soft constraints. We apply it to induce hypernymy relations by training with is-a pairs. We also present an augmented variant of SPON that can generalize type information learned for in-vocabulary terms to previously unseen ones. An extensive ev… ▽ More This paper introduces Strict Partial Order Networks (SPON), a novel neural network architecture designed to enforce asymmetry and transitive properties as soft constraints. We apply it to induce hypernymy relations by training with is-a pairs. We also present an augmented variant of SPON that can generalize type information learned for in-vocabulary terms to previously unseen ones. An extensive evaluation over eleven benchmarks across different tasks shows that SPON consistently either outperforms or attains the state of the art on all but one of these benchmarks. △ Less

Submitted 22 November, 2019; v1 submitted 23 September, 2019; originally announced September 2019.

Comments: 8 pages

Journal ref: AAAI 2020

Interictal intracranial EEG for predicting surgical success: the importance of space and time

Authors: Yujiang Wang, Nishant Sinha, Gabrielle M. Schroeder, Sriharsha Ramaraju, Andrew W. McEvoy, Anna Miserocchi, Jane de Tisi, Fahmida A. Chowdhury, Beate Diehl, John S. Duncan, Peter N. Taylor

Abstract: Predicting post-operative seizure freedom using functional correlation networks derived from interictal intracranial EEG has shown some success. However, there are important challenges to consider. 1: electrodes physically closer to each other naturally tend to be more correlated causing a spatial bias. 2: implantation location and number of electrodes differ between patients, making cross-subject… ▽ More Predicting post-operative seizure freedom using functional correlation networks derived from interictal intracranial EEG has shown some success. However, there are important challenges to consider. 1: electrodes physically closer to each other naturally tend to be more correlated causing a spatial bias. 2: implantation location and number of electrodes differ between patients, making cross-subject comparisons difficult. 3: functional correlation networks can vary over time but are currently assumed as static. In this study we address these three substantial challenges using intracranial EEG data from 55 patients with intractable focal epilepsy. Patients additionally underwent preoperative MR imaging, intra-operative CT, and post-operative MRI allowing accurate localisation of electrodes and delineation of removed tissue. We show that normalising for spatial proximity between nearby electrodes improves prediction of post-surgery seizure outcomes. Moreover, patients with more extensive electrode coverage were more likely to have their outcome predicted correctly (ROC-AUC >0.9, p<<0.05), but not necessarily more likely to have a better outcome. Finally, our predictions are robust regardless of the time segment. Future studies should account for the spatial proximity of electrodes in functional network construction to improve prediction of post-surgical seizure outcomes. Greater coverage of both removed and spared tissue allows for predictions with higher accuracy. △ Less

Submitted 17 September, 2019; originally announced September 2019.

Journal ref: Epilepsia 61 (2020) 1417-1426

SLSNet: Skin lesion segmentation using a lightweight generative adversarial network

Authors: Md. Mostafa Kamal Sarker, Hatem A. Rashwan, Farhan Akram, Vivek Kumar Singh, Syeda Furruka Banu, Forhad U H Chowdhury, Kabir Ahmed Choudhury, Sylvie Chambon, Petia Radeva, Domenec Puig, Mohamed Abdel-Nasser

Abstract: The determination of precise skin lesion boundaries in dermoscopic images using automated methods faces many challenges, most importantly, the presence of hair, inconspicuous lesion edges and low contrast in dermoscopic images, and variability in the color, texture and shapes of skin lesions. Existing deep learning-based skin lesion segmentation algorithms are expensive in terms of computational t… ▽ More The determination of precise skin lesion boundaries in dermoscopic images using automated methods faces many challenges, most importantly, the presence of hair, inconspicuous lesion edges and low contrast in dermoscopic images, and variability in the color, texture and shapes of skin lesions. Existing deep learning-based skin lesion segmentation algorithms are expensive in terms of computational time and memory. Consequently, running such segmentation algorithms requires a powerful GPU and high bandwidth memory, which are not available in dermoscopy devices. Thus, this article aims to achieve precise skin lesion segmentation with minimum resources: a lightweight, efficient generative adversarial network (GAN) model called SLSNet, which combines 1-D kernel factorized networks, position and channel attention, and multiscale aggregation mechanisms with a GAN model. The 1-D kernel factorized network reduces the computational cost of 2D filtering. The position and channel attention modules enhance the discriminative ability between the lesion and non-lesion feature representations in spatial and channel dimensions, respectively. A multiscale block is also used to aggregate the coarse-to-fine features of input skin images and reduce the effect of the artifacts. SLSNet is evaluated on two publicly available datasets: ISBI 2017 and the ISIC 2018. Although SLSNet has only 2.35 million parameters, the experimental results demonstrate that it achieves segmentation results on a par with the state-of-the-art skin lesion segmentation methods with an accuracy of 97.61%, and Dice and Jaccard similarity coefficients of 90.63% and 81.98%, respectively. SLSNet can run at more than 110 frames per second (FPS) in a single GTX1080Ti GPU, which is faster than well-known deep learning-based image segmentation models, such as FCN. Therefore, SLSNet can be used for practical dermoscopic applications. △ Less

Submitted 17 June, 2021; v1 submitted 1 July, 2019; originally announced July 2019.

Comments: Accepted in Expert Systems with Applications

An Efficient Approach for Super and Nested Term Indexing and Retrieval

Authors: Md Faisal Mahbub Chowdhury, Robert Farrell

Abstract: This paper describes a new approach, called Terminological Bucket Indexing (TBI), for efficient indexing and retrieval of both nested and super terms using a single method. We propose a hybrid data structure for facilitating faster indexing building. An evaluation of our approach with respect to widely used existing approaches on several publicly available dataset is provided. Compared to Trie bas… ▽ More This paper describes a new approach, called Terminological Bucket Indexing (TBI), for efficient indexing and retrieval of both nested and super terms using a single method. We propose a hybrid data structure for facilitating faster indexing building. An evaluation of our approach with respect to widely used existing approaches on several publicly available dataset is provided. Compared to Trie based approaches, TBI provides comparable performance on nested term retrieval and far superior performance on super term retrieval. Compared to traditional hash table, TBI needs 80\% less time for indexing. △ Less

Submitted 23 May, 2019; originally announced May 2019.

Comments: 6 pages including references

Topics of Concern: Identifying User Issues in Reviews of IoT Apps and Devices

Authors: Andrew Truelove, Farah Naz Chowdhury, Omprakash Gnawali, Mohammad Amin Alipour

Abstract: Internet of Things (IoT) systems are bundles of networked sensors and actuators that are deployed in an environment and act upon the sensory data that they receive. These systems, especially consumer electronics, have two main cooperating components: a device and a mobile app. The unique combination of hardware and software in IoT systems presents challenges that are lesser known to mainstream sof… ▽ More Internet of Things (IoT) systems are bundles of networked sensors and actuators that are deployed in an environment and act upon the sensory data that they receive. These systems, especially consumer electronics, have two main cooperating components: a device and a mobile app. The unique combination of hardware and software in IoT systems presents challenges that are lesser known to mainstream software developers. They might require innovative solutions to support the development and integration of such systems. In this paper, we analyze more than 90,000 reviews of ten IoT devices and their corresponding apps and extract the issues that users encountered while using these systems. Our results indicate that issues with connectivity, timing, and updates are particularly prevalent in the reviews. Our results call for a new software-hardware development framework to assist the development of reliable IoT systems. △ Less

Submitted 29 March, 2019; v1 submitted 17 February, 2019; originally announced February 2019.

Comments: 1st International Workshop on Software Engineering Research & Practices for the Internet of Things (SERP4IoT 2019)

Yttrium-substituted Mg-Zn ferrites: correlation of physical properties with Yttrium content

Authors: M. A. Ali, M. N. I. Khan, F. -U. -Z. Chowdhury, M. M. Hossain, A. K. M. Akhter Hossain, R. Rashid, A. Nahar, S. M. Hoque, M. A. Matin, M. M. Uddin

Abstract: Yttrium (Y) substituted Mg Zn ferrites with the compositions of Mg0.5Zn0.5YxFe2 xO4 have been synthesized by conventional standard ceramic technique. The effect of Y3+ substitution on the structural, electrical, dielectric and magnetic properties of Mg Zn ferrites has been studied. Yttrium (Y) substituted Mg Zn ferrites with the compositions of Mg0.5Zn0.5YxFe2 xO4 have been synthesized by conventional standard ceramic technique. The effect of Y3+ substitution on the structural, electrical, dielectric and magnetic properties of Mg Zn ferrites has been studied. △ Less

Submitted 7 July, 2019; v1 submitted 15 December, 2018; originally announced December 2018.

Comments: 25 pages, 8 figure

Journal ref: J Mater Sci: Mater Electron (2019)

Centralized adaptive traffic control strategy design across multiple intersections based on vehicle path flows: An approximated Lagrangian decomposition approach

Authors: Peirong Wang, Pengfei Li, Farzana Chowdhury, Li Zhang

Abstract: In this paper, we first present a centralized traffic control model based on the emerging dynamic path flows. This new model in essence views the whole target network as one integral piece in which traffic propagates based on traffic flow dynamics, vehicle paths, and traffic control. In light of this centralized traffic control concept, most requirements for the existing traffic control coordinati… ▽ More In this paper, we first present a centralized traffic control model based on the emerging dynamic path flows. This new model in essence views the whole target network as one integral piece in which traffic propagates based on traffic flow dynamics, vehicle paths, and traffic control. In light of this centralized traffic control concept, most requirements for the existing traffic control coordination will be dropped in the optimal traffic control operations, such as the common cycle length or offsets. Instead, the optimal traffic control strategy across intersections will be highly adaptive over time to minimize the total travel time and it can also prevent any short-term traffic congestions according to the space-time characteristics of vehicle path flows. A mixed integer linear programming (MILP) formulation is then presented to model the propagations of path flows given a centralized traffic control strategy. Secondly, a new approximated Lagrangian decomposition framework is presented. Solving the proposed MILP model with the traditional Lagrangian decomposition approach faces the leveraging challenge between the problem tractability after decomposition and computing complexity. To address this issue, we propose an approximate Lagrangian decomposition framework in which the target problem is reformulated into an approximated problem by constructing its complicating attributes into the objective function instead of in constraints to reduce the computing complexity. The computing efficiency loss of the objective function can be mitigated via multiple advanced computing techniques. With the proposed approximated approach, the upper bound and tight lower bound can be obtained via two customized dynamic network loading models. In the end, one demonstrative and one real-world example are illustrated to show the validness, robustness, and scalability of the new approach. △ Less

Submitted 10 November, 2018; originally announced November 2018.

SLSDeep: Skin Lesion Segmentation Based on Dilated Residual and Pyramid Pooling Networks

Authors: Md. Mostafa Kamal Sarker, Hatem A. Rashwan, Farhan Akram, Syeda Furruka Banu, Adel Saleh, Vivek Kumar Singh, Forhad U H Chowdhury, Saddam Abdulwahab, Santiago Romani, Petia Radeva, Domenec Puig

Abstract: Skin lesion segmentation (SLS) in dermoscopic images is a crucial task for automated diagnosis of melanoma. In this paper, we present a robust deep learning SLS model, so-called SLSDeep, which is represented as an encoder-decoder network. The encoder network is constructed by dilated residual layers, in turn, a pyramid pooling network followed by three convolution layers is used for the decoder. U… ▽ More Skin lesion segmentation (SLS) in dermoscopic images is a crucial task for automated diagnosis of melanoma. In this paper, we present a robust deep learning SLS model, so-called SLSDeep, which is represented as an encoder-decoder network. The encoder network is constructed by dilated residual layers, in turn, a pyramid pooling network followed by three convolution layers is used for the decoder. Unlike the traditional methods employing a cross-entropy loss, we investigated a loss function by combining both Negative Log Likelihood (NLL) and End Point Error (EPE) to accurately segment the melanoma regions with sharp boundaries. The robustness of the proposed model was evaluated on two public databases: ISBI 2016 and 2017 for skin lesion analysis towards melanoma detection challenge. The proposed model outperforms the state-of-the-art methods in terms of segmentation accuracy. Moreover, it is capable to segment more than $100$ images of size 384x384 per second on a recent GPU. △ Less

Submitted 30 May, 2018; v1 submitted 25 May, 2018; originally announced May 2018.

Comments: Accepted in MICCAI 2018, 9 pages

A Study on Passage Re-ranking in Embedding based Unsupervised Semantic Search

Authors: Md Faisal Mahbub Chowdhury, Vijil Chenthamarakshan, Rishav Chakravarti, Alfio M. Gliozzo

Abstract: State of the art approaches for (embedding based) unsupervised semantic search exploits either compositional similarity (of a query and a passage) or pair-wise word (or term) similarity (from the query and the passage). By design, word based approaches do not incorporate similarity in the larger context (query/passage), while compositional similarity based approaches are usually unable to take adv… ▽ More State of the art approaches for (embedding based) unsupervised semantic search exploits either compositional similarity (of a query and a passage) or pair-wise word (or term) similarity (from the query and the passage). By design, word based approaches do not incorporate similarity in the larger context (query/passage), while compositional similarity based approaches are usually unable to take advantage of the most important cues in the context. In this paper we propose a new compositional similarity based approach, called variable centroid vector (VCVB), that tries to address both of these limitations. We also presents results using a different type of compositional similarity based approach by exploiting universal sentence embedding. We provide empirical evaluation on two different benchmarks. △ Less

Submitted 13 March, 2019; v1 submitted 21 April, 2018; originally announced April 2018.

Comments: Fixed latex compiling issues