Showing 1–50 of 56 results for author: Sumedha

Keyword spotting -- Detecting commands in speech using deep learning

Authors: Sumedha Rai, Tong Li, Bella Lyu

Abstract: Speech recognition has become an important task in the development of machine learning and artificial intelligence. In this study, we explore the important task of keyword spotting using speech recognition machine learning and deep learning techniques. We implement feature engineering by converting raw waveforms to Mel Frequency Cepstral Coefficients (MFCCs), which we use as inputs to our models.… ▽ More Speech recognition has become an important task in the development of machine learning and artificial intelligence. In this study, we explore the important task of keyword spotting using speech recognition machine learning and deep learning techniques. We implement feature engineering by converting raw waveforms to Mel Frequency Cepstral Coefficients (MFCCs), which we use as inputs to our models. We experiment with several different algorithms such as Hidden Markov Model with Gaussian Mixture, Convolutional Neural Networks and variants of Recurrent Neural Networks including Long Short-Term Memory and the Attention mechanism. In our experiments, RNN with BiLSTM and Attention achieves the best performance with an accuracy of 93.9 % △ Less

Submitted 9 December, 2023; originally announced December 2023.

Inverse transitions and disappearance of the λ-line in the asymmetric random field Ising and Blume-Capel models

Authors: Santanu Das, Sumedha

Abstract: We report on reentrance in the random field Ising and Blume-Capel models, induced by an asymmetric bimodal random field distribution. The conventional continuous line of transitions between the paramagnetic and ferromagnetic phases, the λ-line, is wiped away by the asymmetry. The phase diagram, then, consists of only first order transition lines that always end at ordered critical points. We find… ▽ More We report on reentrance in the random field Ising and Blume-Capel models, induced by an asymmetric bimodal random field distribution. The conventional continuous line of transitions between the paramagnetic and ferromagnetic phases, the λ-line, is wiped away by the asymmetry. The phase diagram, then, consists of only first order transition lines that always end at ordered critical points. We find that while for symmetric random field distributions there was no reentrance, the asymmetry in the random field results in a range of temperatures for which magnetisation shows reentrance. While this does not give rise to an inverse transition in the Ising model, for the Blume-Capel model, however, there is a line of first order inverse phase transitions that ends at an inverse ordered critical point. We show that the location of the inverse transitions can be inferred from the ground state phase diagram of the model. △ Less

Submitted 7 October, 2023; originally announced October 2023.

Comments: 6 pages, 7 figures, Accepted for publication in Phys. Rev. E (Letter)

Preparing for Gaia Searches for Optical Counterparts of Gravitational Wave Events during O4

Authors: Sumedha Biswas, Zuzanna Kostrzewa-Rutkowska, Peter G. Jonker, Paul Vreeswijk, Deepak Eappachen, Paul J. Groot, Simon Hodgkin, Abdullah Yoldas, Guy Rixon, Diana Harrison, M. van Leeuwen, Dafydd Evans

Abstract: The discovery of gravitational wave (GW) events and the detection of electromagnetic counterparts from GW170817 has started the era of multimessenger GW astronomy.The field has been develo** rapidly and in this paper,we discuss the preparation for detecting these events with the ESA Gaia satellite,during the 4th observing run of the LIGO-Virgo-KAGRA (LVK) collaboration that has started on May 24… ▽ More The discovery of gravitational wave (GW) events and the detection of electromagnetic counterparts from GW170817 has started the era of multimessenger GW astronomy.The field has been develo** rapidly and in this paper,we discuss the preparation for detecting these events with the ESA Gaia satellite,during the 4th observing run of the LIGO-Virgo-KAGRA (LVK) collaboration that has started on May 24,2023. Gaia is contributing to the search for GW counterparts by a new transient detection pipeline called GaiaX. In GaiaX, a new source appearing in the field of view of only one of the two telescopes on-board Gaia is sufficient to send out an alert on the possible detection of a new transient. Ahead of O4, an experiment was conducted over a period of about two months. During the two weeks around New Moon in this period of time, the MeerLICHT (ML) telescope located in South Africa tried (weather permitting) to observe the same region of the sky as Gaia within 10 minutes. Any GaiaX detected transient was published publicly. ML and Gaia have similar limiting magnitudes for typical seeing conditions at ML. At the end of the experiment, we had 11861 GaiaX candidate transients and 15806 ML candidate transients, which we further analysed and the results of which are presented in this paper. Finally, we discuss the possibility and capabilities of Gaia contributing to the search for electromagnetic counterparts of gravitational wave events during O4 through the GaiaX detection and alert procedure. △ Less

Submitted 21 August, 2023; v1 submitted 11 July, 2023; originally announced July 2023.

Comments: 13 pages, 9 figures; Accepted for publication in MNRAS;

Decentralized Data Governance as Part of a Data Mesh Platform: Concepts and Approaches

Authors: Arif Wider, Sumedha Verma, Atif Akhtar

Abstract: Data mesh is a socio-technical approach to decentralized analytics data management. To manage this decentralization efficiently, data mesh relies on automation provided by a self-service data infrastructure platform. A key aspect of this platform is to enable decentralized data governance. Because data mesh is a young approach, there is a lack of coherence in how data mesh concepts are interpreted… ▽ More Data mesh is a socio-technical approach to decentralized analytics data management. To manage this decentralization efficiently, data mesh relies on automation provided by a self-service data infrastructure platform. A key aspect of this platform is to enable decentralized data governance. Because data mesh is a young approach, there is a lack of coherence in how data mesh concepts are interpreted in the industry, and almost no work on how a data mesh platform facilitates governance. This paper presents a conceptual model of key data mesh concepts and discusses different approaches to drive governance through platform means. The insights presented are drawn from concrete experiences of implementing a fully-functional data mesh platform that can be used as a reference on how to approach data mesh platform development. △ Less

Submitted 5 July, 2023; originally announced July 2023.

Comments: 30th International Conference on Web Services, July 2-8, 2023, Chicago, USA

Critical behaviour near critical end points and tricritical points in disordered spin-1 ferromagnets

Authors: Soheli Mukherjee, Sumedha

Abstract: Critical end points and tricritical points are multicritical points that separate lines of continuous transitions from lines of first order transitions in the phase diagram of many systems. In models like the spin-1 disordered Blume-Capel model and the repulsive Blume-Emery-Griffiths model, the tricritical point splits into a critical end point and a bicritical end point with an increase in disord… ▽ More Critical end points and tricritical points are multicritical points that separate lines of continuous transitions from lines of first order transitions in the phase diagram of many systems. In models like the spin-1 disordered Blume-Capel model and the repulsive Blume-Emery-Griffiths model, the tricritical point splits into a critical end point and a bicritical end point with an increase in disorder and repulsive coupling strength respectively. In order to make a distinction between these two multicritical points, we investigate and contrast the behaviour of the first order phase boundary and the co-existence diameter around them. △ Less

Submitted 14 May, 2023; originally announced May 2023.

Comments: 23 pages, 28 figures

Journal ref: Physica A 624 (2023) 128905

Augmentation by Counterfactual Explanation -- Fixing an Overconfident Classifier

Authors: Sumedha Singla, Nihal Murali, Forough Arabshahi, Sofia Triantafyllou, Kayhan Batmanghelich

Abstract: A highly accurate but overconfident model is ill-suited for deployment in critical applications such as healthcare and autonomous driving. The classification outcome should reflect a high uncertainty on ambiguous in-distribution samples that lie close to the decision boundary. The model should also refrain from making overconfident decisions on samples that lie far outside its training distributio… ▽ More A highly accurate but overconfident model is ill-suited for deployment in critical applications such as healthcare and autonomous driving. The classification outcome should reflect a high uncertainty on ambiguous in-distribution samples that lie close to the decision boundary. The model should also refrain from making overconfident decisions on samples that lie far outside its training distribution, far-out-of-distribution (far-OOD), or on unseen samples from novel classes that lie near its training distribution (near-OOD). This paper proposes an application of counterfactual explanations in fixing an over-confident classifier. Specifically, we propose to fine-tune a given pre-trained classifier using augmentations from a counterfactual explainer (ACE) to fix its uncertainty characteristics while retaining its predictive performance. We perform extensive experiments with detecting far-OOD, near-OOD, and ambiguous samples. Our empirical results show that the revised model have improved uncertainty measures, and its performance is competitive to the state-of-the-art methods. △ Less

Submitted 21 October, 2022; originally announced October 2022.

Comments: Accepted in WACV 2023

Deep Learning for Medical Imaging From Diagnosis Prediction to its Counterfactual Explanation

Authors: Sumedha Singla

Abstract: Deep neural networks (DNN) have achieved unprecedented performance in computer-vision tasks almost ubiquitously in business, technology, and science. While substantial efforts are made to engineer highly accurate architectures and provide usable model explanations, most state-of-the-art approaches are first designed for natural vision and then translated to the medical domain. This dissertation se… ▽ More Deep neural networks (DNN) have achieved unprecedented performance in computer-vision tasks almost ubiquitously in business, technology, and science. While substantial efforts are made to engineer highly accurate architectures and provide usable model explanations, most state-of-the-art approaches are first designed for natural vision and then translated to the medical domain. This dissertation seeks to address this gap by proposing novel architectures that integrate the domain-specific constraints of medical imaging into the DNN model and explanation design. △ Less

Submitted 7 September, 2022; originally announced September 2022.

Comments: PhD thesis

Multiple transitions in an infinite range p-spin random-crystal field Blume Capel model

Authors: Santanu Das, Sumedha

Abstract: We study a $p$-spin model with ferromagnetic coupling and quenched random-crystal fields for $p \ge 3$ for spin-1 systems. We find that the model has lines of first order transitions at finite temperature $(T)$ for all $p \ge 3$. For bimodal distribution of the random-crystal field these lines meet at a \emph{triple point} for weak strength of the crystal field $(Δ)$. Beyond a critical strength of… ▽ More We study a $p$-spin model with ferromagnetic coupling and quenched random-crystal fields for $p \ge 3$ for spin-1 systems. We find that the model has lines of first order transitions at finite temperature $(T)$ for all $p \ge 3$. For bimodal distribution of the random-crystal field these lines meet at a \emph{triple point} for weak strength of the crystal field $(Δ)$. Beyond a critical strength of $Δ$, they do not meet and one of the lines ends at a \emph{critical point} $(T_c)$. Interestingly, we find that on increasing $T$ from $T_c$ kee** other parameters fixed, the system undergoes one more transition which is first order in its character. The system thus exhibits a Gardner like transition for a range of parameters for all finite $p \ge 3$. For $p \to \infty$ the model behaves differently and there is only one random first order transition at $T = 0$. △ Less

Submitted 12 May, 2022; originally announced May 2022.

Journal ref: Phys. Rev. E 106, 034134 (2022)

Phase transitions in the Blume-Capel model with trimodal and Gaussian random fields

Authors: Soheli Mukherjee, Sumedha

Abstract: We study the effect of different symmetric random field distributions: trimodal and Gaussian on the phase diagram of the infinite range Blume-Capel model. For the trimodal random field, the model has a very rich phase diagram. We find three new ordered phases, multicritical points like tricritical point (TCP), bicritical end point (BEP), critical end point (CEP) along with some multi-phase coexist… ▽ More We study the effect of different symmetric random field distributions: trimodal and Gaussian on the phase diagram of the infinite range Blume-Capel model. For the trimodal random field, the model has a very rich phase diagram. We find three new ordered phases, multicritical points like tricritical point (TCP), bicritical end point (BEP), critical end point (CEP) along with some multi-phase coexistence points. We also find re-entrance at low temperatures for some values of the parameters. On the other hand for the Gaussian distribution the phase diagram consists of a continuous line of transition followed by a first order transition line, meeting at a TCP. The TCP vanishes for higher strength of the random field. In contrast to the trimodal case, in Gaussian case no new phase emerges. △ Less

Submitted 31 May, 2022; v1 submitted 10 March, 2022; originally announced March 2022.

Comments: 25 pages, 19 figures. Version accepted for publication in Journal of Statistical Physics

Journal ref: Journal of Statistical Physics volume 188, Article number: 22 (2022)

Phase transitions in XY models with randomly oriented crystal fields

Authors: Sumedha, M. Barma

Abstract: We obtain a representation of the free energy of an XY model on a fully connected graph with spins subjected to a random crystal field of strength $D$ and with random orientation $α$. Results are obtained for an arbitrary probability distribution of the disorder using large deviation theory, for any $D$. We show that the critical temperature is insensitive to the nature and strength of the distrib… ▽ More We obtain a representation of the free energy of an XY model on a fully connected graph with spins subjected to a random crystal field of strength $D$ and with random orientation $α$. Results are obtained for an arbitrary probability distribution of the disorder using large deviation theory, for any $D$. We show that the critical temperature is insensitive to the nature and strength of the distribution $p(α)$, for a large family of distributions which includes quadriperiodic distributions, with $p(α)=p(α+\fracπ{2})$, which includes the uniform and symmetric bimodal distributions. The specific heat vanishes as temperature $T \rightarrow 0$ if $D$ is infinite, but approaches a constant if $D$ is finite. We also studied the effect of asymmetry on a bimodal distribution of the orientation of the random crystal field and obtained the phase diagram comprising four phases: a mixed phase (in which spins are canted at angles which depend on the degree of disorder), an $x$-Ising phase, a $y$-Ising phase and a paramagnetic phase, all of which meet at a tetra-critical point. The canted mixed phase is present for all finite $D$, but vanishes when $D \rightarrow \infty$. △ Less

Submitted 24 January, 2022; originally announced January 2022.

Comments: 14 pages, 9 figures

Journal ref: Phys. Rev. E 105, 024111 (2022)

Hysteresis and return point memory in the random field Blume Capel model

Authors: B. E. Aldrin, Abdul Khaleque, Sumedha

Abstract: We study the zero temperature steady state of the random field Blume Capel model with spin-flip Glauber dynamics on a random regular graph. The magnetization m as a function of the external field H is observed to have double hysteresis loops with a return point memory. We also solve the model on a Bethe lattice in the approximation that the spin relaxation dynamics is abelian and find good agreeme… ▽ More We study the zero temperature steady state of the random field Blume Capel model with spin-flip Glauber dynamics on a random regular graph. The magnetization m as a function of the external field H is observed to have double hysteresis loops with a return point memory. We also solve the model on a Bethe lattice in the approximation that the spin relaxation dynamics is abelian and find good agreement between simulations on random regular graphs and Bethe lattice calculations for negative values of H. △ Less

Submitted 29 June, 2022; v1 submitted 24 November, 2021; originally announced November 2021.

Comments: version accepted in Phys. Rev. E

Journal ref: Phys. Rev. E 106 014129(2022)

Approximating Sparsest Cut in Low-Treewidth Graphs via Combinatorial Diameter

Authors: Parinya Chalermsook, Matthias Kaul, Matthias Mnich, Joachim Spoerhase, Sumedha Uniyal, Daniel Vaz

Abstract: The fundamental sparsest cut problem takes as input a graph $G$ together with the edge costs and demands, and seeks a cut that minimizes the ratio between the costs and demands across the cuts. For $n$-node graphs~$G$ of treewidth~$k$, \chlamtac, Krauthgamer, and Raghavendra (APPROX 2010) presented an algorithm that yields a factor-$2^{2^k}$ approximation in time… ▽ More The fundamental sparsest cut problem takes as input a graph $G$ together with the edge costs and demands, and seeks a cut that minimizes the ratio between the costs and demands across the cuts. For $n$-node graphs~$G$ of treewidth~$k$, \chlamtac, Krauthgamer, and Raghavendra (APPROX 2010) presented an algorithm that yields a factor-$2^{2^k}$ approximation in time $2^{O(k)} \cdot \operatorname{poly}(n)$. Later, Gupta, Talwar and Witmer (STOC 2013) showed how to obtain a $2$-approximation algorithm with a blown-up run time of $n^{O(k)}$. An intriguing open question is whether one can simultaneously achieve the best out of the aforementioned results, that is, a factor-$2$ approximation in time $2^{O(k)} \cdot \operatorname{poly}(n)$. In this paper, we make significant progress towards this goal, via the following results: (i) A factor-$O(k^2)$ approximation that runs in time $2^{O(k)} \cdot \operatorname{poly}(n)$, directly improving the work of Chlamtáč et al. while kee** the run time single-exponential in $k$. (ii) For any $\varepsilon>0$, a factor-$O(1/\varepsilon^2)$ approximation whose run time is $2^{O(k^{1+\varepsilon}/\varepsilon)} \cdot \operatorname{poly}(n)$, implying a constant-factor approximation whose run time is nearly single-exponential in $k$ and a factor-$O(\log^2 k)$ approximation in time $k^{O(k)} \cdot \operatorname{poly}(n)$. Key to these results is a new measure of a tree decomposition that we call combinatorial diameter, which may be of independent interest. △ Less

Submitted 11 November, 2021; originally announced November 2021.

Comments: 15 pages, 3 figures

Rejection-free cluster Wang-Landau algorithm for hard-core lattice gases

Authors: Asweel Ahmed A. Jaleel, Jetin E. Thomas, Dipanjan Mandal, Sumedha, R. Rajesh

Abstract: We introduce a rejection-free, flat histogram, cluster algorithm to determine the density of states of hard-core lattice gases. We show that the algorithm is able to efficiently sample low entropy states that are usually difficult to access, even when the excluded volume per particle is large. The algorithm is based on simultaneously evaporating all the particles in a strip and reoccupying these s… ▽ More We introduce a rejection-free, flat histogram, cluster algorithm to determine the density of states of hard-core lattice gases. We show that the algorithm is able to efficiently sample low entropy states that are usually difficult to access, even when the excluded volume per particle is large. The algorithm is based on simultaneously evaporating all the particles in a strip and reoccupying these sites with a new appropriately chosen configuration. We implement the algorithm for the particular case of the hard-core lattice gas in which the first k next-nearest neighbors of a particle are excluded from being occupied. It is shown that the algorithm is able to reproduce the known results for k = 1,2,3 both on the square and cubic lattices. We also show that, in comparison, the corresponding flat histogram algorithms with either local moves or unbiased cluster moves are less accurate and do not converge as the system size increases. △ Less

Submitted 29 October, 2021; v1 submitted 3 August, 2021; originally announced August 2021.

Comments: 19 Pages, 25 figures

Journal ref: Phys. Rev. E 104, 045310 (2021)

Using Causal Analysis for Conceptual Deep Learning Explanation

Authors: Sumedha Singla, Stephen Wallace, Sofia Triantafillou, Kayhan Batmanghelich

Abstract: Model explainability is essential for the creation of trustworthy Machine Learning models in healthcare. An ideal explanation resembles the decision-making process of a domain expert and is expressed using concepts or terminology that is meaningful to the clinicians. To provide such an explanation, we first associate the hidden units of the classifier to clinically relevant concepts. We take advan… ▽ More Model explainability is essential for the creation of trustworthy Machine Learning models in healthcare. An ideal explanation resembles the decision-making process of a domain expert and is expressed using concepts or terminology that is meaningful to the clinicians. To provide such an explanation, we first associate the hidden units of the classifier to clinically relevant concepts. We take advantage of radiology reports accompanying the chest X-ray images to define concepts. We discover sparse associations between concepts and hidden units using a linear sparse logistic regression. To ensure that the identified units truly influence the classifier's outcome, we adopt tools from Causal Inference literature and, more specifically, mediation analysis through counterfactual interventions. Finally, we construct a low-depth decision tree to translate all the discovered concepts into a straightforward decision rule, expressed to the radiologist. We evaluated our approach on a large chest x-ray dataset, where our model produces a global explanation consistent with clinical knowledge. △ Less

Submitted 9 July, 2021; originally announced July 2021.

Comments: 10 pages, 6 figures

Solution of the random field $XY$ magnet on a fully connected graph

Authors: Sumedha, Mustansir Barma

Abstract: We use large deviation theory to obtain the free energy of the XY model on a fully connected graph on each site of which there is a randomly oriented field of magnitude $h$. The phase diagram is obtained for two symmetric distributions of the random orientations: (a) a uniform distribution and (b) a distribution with cubic symmetry. In both cases, the disorder-averaged ordered state reflects the s… ▽ More We use large deviation theory to obtain the free energy of the XY model on a fully connected graph on each site of which there is a randomly oriented field of magnitude $h$. The phase diagram is obtained for two symmetric distributions of the random orientations: (a) a uniform distribution and (b) a distribution with cubic symmetry. In both cases, the disorder-averaged ordered state reflects the symmetry of the underlying distribution. The phase boundary has a multicritical point which separates a locus of continuous transitions (for small values of $h$) from a locus of first order transitions (for large $h$). The free energy is a function of a single variable in case (a) and a function of two variables in case (b), leading to different characters of the multicritical points in the two cases. We find that the locus of continuous transitions is given by the same equation for a family of quadriperiodic distributions, which includes the distributions (a) and (b). However, the location of the multicritical point and the nature of ordered state depend on the form of the distribution. The disorder-averaged ground state energy is found exactly, and the specific heat is shown to approach a constant as temperature approaches zero. △ Less

Submitted 24 January, 2022; v1 submitted 14 April, 2021; originally announced April 2021.

Comments: replaced with the version accepted for publication in J. Phys. A : Math. Theor

Journal ref: J. Phys. A: Math. Theor. 55 095001 (2022)

Self-Supervised Vessel Enhancement Using Flow-Based Consistencies

Authors: Rohit Jena, Sumedha Singla, Kayhan Batmanghelich

Abstract: Vessel segmentation is an essential task in many clinical applications. Although supervised methods have achieved state-of-art performance, acquiring expert annotation is laborious and mostly limited for two-dimensional datasets with a small sample size. On the contrary, unsupervised methods rely on handcrafted features to detect tube-like structures such as vessels. However, those methods require… ▽ More Vessel segmentation is an essential task in many clinical applications. Although supervised methods have achieved state-of-art performance, acquiring expert annotation is laborious and mostly limited for two-dimensional datasets with a small sample size. On the contrary, unsupervised methods rely on handcrafted features to detect tube-like structures such as vessels. However, those methods require complex pipelines involving several hyper-parameters and design choices rendering the procedure sensitive, dataset-specific, and not generalizable. We propose a self-supervised method with a limited number of hyper-parameters that is generalizable across modalities. Our method uses tube-like structure properties, such as connectivity, profile consistency, and bifurcation, to introduce inductive bias into a learning algorithm. To model those properties, we generate a vector field that we refer to as a flow. Our experiments on various public datasets in 2D and 3D show that our method performs better than unsupervised methods while learning useful transferable features from unlabeled data. Unlike generic self-supervised methods, the learned features learn vessel-relevant features that are transferable for supervised approaches, which is essential when the number of annotated data is limited. △ Less

Submitted 22 July, 2021; v1 submitted 13 January, 2021; originally announced January 2021.

Comments: Early accept at MICCAI 2021

Explaining the Black-box Smoothly- A Counterfactual Approach

Authors: Sumedha Singla, Motahhare Eslami, Brian Pollack, Stephen Wallace, Kayhan Batmanghelich

Abstract: We propose a BlackBox Counterfactual Explainer, designed to explain image classification models for medical applications. Classical approaches (e.g., saliency maps) that assess feature importance do not explain "how" imaging features in important anatomical regions are relevant to the classification decision. Our framework explains the decision for a target class by gradually "exaggerating" the se… ▽ More We propose a BlackBox Counterfactual Explainer, designed to explain image classification models for medical applications. Classical approaches (e.g., saliency maps) that assess feature importance do not explain "how" imaging features in important anatomical regions are relevant to the classification decision. Our framework explains the decision for a target class by gradually "exaggerating" the semantic effect of the class in a query image. We adopted a Generative Adversarial Network (GAN) to generate a progressive set of perturbations to a query image, such that the classification decision changes from its original class to its negation. We used counterfactual explanations from our framework to audit a classifier trained on a chest x-ray dataset with multiple labels. We proposed clinically-relevant quantitative metrics such as cardiothoracic ratio and the score of a healthy costophrenic recess to evaluate our explanations. We conducted a human-grounded experiment with diagnostic radiology residents to compare different styles of explanations (no explanation, saliency map, cycleGAN explanation, and our counterfactual explanation) by evaluating different aspects of explanations: (1) understandability, (2) classifier's decision justification, (3) visual quality, (d) identity preservation, and (5) overall helpfulness of an explanation to the users. Our results show that our counterfactual explanation was the only explanation method that significantly improved the users' understanding of the classifier's decision compared to the no-explanation baseline. Our metrics established a benchmark for evaluating model explanation methods in medical images. Our explanations revealed that the classifier relied on clinically relevant radiographic features for its diagnostic decisions, thus making its decision-making process more transparent to the end-user. △ Less

Submitted 18 November, 2022; v1 submitted 11 January, 2021; originally announced January 2021.

Comments: Preprint Accepted in Medical image Analysis journal

Phase diagram of the repulsive Blume-Emery-Griffiths model in the presence of external magnetic field on a complete graph

Authors: Soheli Mukherjee, Raj Kumar Sadhu, Sumedha

Abstract: For the repulsive Blume-Emery-Griffiths model the phase diagram in the space of three fields, temperature (T), crystal field ($Δ$), and magnetic field (H), is computed on a complete graph, in the canonical and microcanonical ensembles. For weak strength of the biquadratic interaction (K), there exists a tricritical point in the phase diagram where three critical lines meet. As K decreases below a… ▽ More For the repulsive Blume-Emery-Griffiths model the phase diagram in the space of three fields, temperature (T), crystal field ($Δ$), and magnetic field (H), is computed on a complete graph, in the canonical and microcanonical ensembles. For weak strength of the biquadratic interaction (K), there exists a tricritical point in the phase diagram where three critical lines meet. As K decreases below a threshold value(which is ensemble dependent), new multicritical points like the critical end point and bicritical end point arise in the (T,$Δ$) plane. For K>-1, we observe that the two critical lines in the H plane and the multicritical points are different in the two ensembles. At K=-1, the two critical lines in the H plane disappear and as K decreases further, there is no phase transition in the H plane. Exactly at K=-1 the two ensembles become equivalent. Beyond that for all K<-1, there are no multicritical points and there is no ensemble inequivalence in the phase diagram. We also study the transition lines in the H plane for positive K i.e. for attractive biquadratic interaction. We find that the transition lines in the H plane are not monotonic in temperature for large positive K. △ Less

Submitted 12 March, 2021; v1 submitted 31 December, 2020; originally announced December 2020.

Comments: 17 pages, 12 figures. Replaced with the version accepted in Journal of Statistical Mechanics: theory and experiment

Journal ref: Journal of Statistical Mechanics: Theory and Experiment, Volume 2021, April 2021

Multi-transversals for Triangles and the Tuza's Conjecture

Authors: Parinya Chalermsook, Samir Khuller, Pattara Sukprasert, Sumedha Uniyal

Abstract: In this paper, we study a primal and dual relationship about triangles: For any graph $G$, let $ν(G)$ be the maximum number of edge-disjoint triangles in $G$, and $τ(G)$ be the minimum subset $F$ of edges such that $G \setminus F$ is triangle-free. It is easy to see that $ν(G) \leq τ(G) \leq 3 ν(G)$, and in fact, this rather obvious inequality holds for a much more general primal-dual relation bet… ▽ More In this paper, we study a primal and dual relationship about triangles: For any graph $G$, let $ν(G)$ be the maximum number of edge-disjoint triangles in $G$, and $τ(G)$ be the minimum subset $F$ of edges such that $G \setminus F$ is triangle-free. It is easy to see that $ν(G) \leq τ(G) \leq 3 ν(G)$, and in fact, this rather obvious inequality holds for a much more general primal-dual relation between $k$-hyper matching and covering in hypergraphs. Tuza conjectured in $1981$ that $τ(G) \leq 2 ν(G)$, and this question has received attention from various groups of researchers in discrete mathematics, settling various special cases such as planar graphs and generalized to bounded maximum average degree graphs, some cases of minor-free graphs, and very dense graphs. Despite these efforts, the conjecture in general graphs has remained wide open for almost four decades. In this paper, we provide a proof of a non-trivial consequence of the conjecture; that is, for every $k \geq 2$, there exist a (multi)-set $F \subseteq E(G): |F| \leq 2k ν(G)$ such that each triangle in $G$ overlaps at least $k$ elements in $F$. Our result can be seen as a strengthened statement of Krivelevich's result on the fractional version of Tuza's conjecture (and we give some examples illustrating this.) The main technical ingredient of our result is a charging argument, that locally identifies edges in $F$ based on a local view of the packing solution. This idea might be useful in further studying the primal-dual relations in general and the Tuza's conjecture in particular. △ Less

Submitted 3 February, 2021; v1 submitted 1 January, 2020; originally announced January 2020.

Comments: Accepted at SODA'20

PTAS for Steiner Tree on Map Graphs

Authors: Jarosław Byrka, Mateusz Lewandowski, Syed Mohammad Meesum, Joachim Spoerhase, Sumedha Uniyal

Abstract: We study the Steiner tree problem on map graphs, which substantially generalize planar graphs as they allow arbitrarily large cliques. We obtain a PTAS for Steiner tree on map graphs, which builds on the result for planar edge weighted instances of Borradaile et al. The Steiner tree problem on map graphs can be casted as a special case of the planar node-weighted Steiner tree problem, for which on… ▽ More We study the Steiner tree problem on map graphs, which substantially generalize planar graphs as they allow arbitrarily large cliques. We obtain a PTAS for Steiner tree on map graphs, which builds on the result for planar edge weighted instances of Borradaile et al. The Steiner tree problem on map graphs can be casted as a special case of the planar node-weighted Steiner tree problem, for which only a 2.4-approximation is known. We prove and use a contraction decomposition theorem for planar node weighted instances. This readily reduces the problem of finding a PTAS for planar node-weighted Steiner tree to finding a spanner, i.e., a constant-factor approximation containing a nearly optimum solution. Finally, we pin-point places where known techniques for constructing such spanner fail on node weighted instances and further progress requires new ideas. △ Less

Submitted 2 December, 2019; originally announced December 2019.

Explanation by Progressive Exaggeration

Authors: Sumedha Singla, Brian Pollack, Junxiang Chen, Kayhan Batmanghelich

Abstract: As machine learning methods see greater adoption and implementation in high stakes applications such as medical image diagnosis, the need for model interpretability and explanation has become more critical. Classical approaches that assess feature importance (e.g. saliency maps) do not explain how and why a particular region of an image is relevant to the prediction. We propose a method that expla… ▽ More As machine learning methods see greater adoption and implementation in high stakes applications such as medical image diagnosis, the need for model interpretability and explanation has become more critical. Classical approaches that assess feature importance (e.g. saliency maps) do not explain how and why a particular region of an image is relevant to the prediction. We propose a method that explains the outcome of a classification black-box by gradually exaggerating the semantic effect of a given class. Given a query input to a classifier, our method produces a progressive set of plausible variations of that query, which gradually changes the posterior probability from its original class to its negation. These counter-factually generated samples preserve features unrelated to the classification decision, such that a user can employ our method as a "tuning knob" to traverse a data manifold while crossing the decision boundary. Our method is model agnostic and only requires the output value and gradient of the predictor with respect to its input. △ Less

Submitted 10 February, 2020; v1 submitted 1 November, 2019; originally announced November 2019.

PARQR: Augmenting the Piazza Online Forum to Better Support Degree Seeking Online Masters Students

Authors: Noah Bilgrien, Roy Finkelberg, Chirag Tailor, India Irish, Girish Murali, Abhishek Mangal, Niklas Gustafsson, Sumedha Raman, Thad Starner, Rosa Arriaga

Abstract: We introduce PARQR, a tool for online education forums that reduces duplicate posts by 40\% in a degree seeking online masters program at a top university. Instead of performing a standard keyword search, PARQR monitors questions as students compose them and continuously suggests relevant posts. In testing, PARQR correctly recommends a relevant post, if one exists, 73.5\% of the time. We discuss P… ▽ More We introduce PARQR, a tool for online education forums that reduces duplicate posts by 40\% in a degree seeking online masters program at a top university. Instead of performing a standard keyword search, PARQR monitors questions as students compose them and continuously suggests relevant posts. In testing, PARQR correctly recommends a relevant post, if one exists, 73.5\% of the time. We discuss PARQR's design, initial experimental results comparing different semesters with and without PARQR, and interviews we conducted with teaching instructors regarding their experience with PARQR. △ Less

Submitted 4 September, 2019; originally announced September 2019.

Emergence of a bicritical end point in the random crystal field Blume-Capel model

Authors: Sumedha, Soheli Mukherjee

Abstract: We obtain the phase diagram for the Blume-Capel model with bimodal distribution for random crystal fields, in the space of three fields: temperature, crystal field and magnetic field. We find that three critical lines meet at a tricritical point, but only for weak disorder. As disorder strength increases there is no tricritical point in the phase diagram. We instead find a bicritical end point, wh… ▽ More We obtain the phase diagram for the Blume-Capel model with bimodal distribution for random crystal fields, in the space of three fields: temperature, crystal field and magnetic field. We find that three critical lines meet at a tricritical point, but only for weak disorder. As disorder strength increases there is no tricritical point in the phase diagram. We instead find a bicritical end point, where only two of the critical lines meet on a first order surface in the H=0 plane. For intermediate strengths of disorder, the phase diagram has critical end points along with the bicritical end point. One needs to look at the phase diagram in the space of three fields to identify various such multicritical points. △ Less

Submitted 29 April, 2020; v1 submitted 15 July, 2019; originally announced July 2019.

Comments: Replaced with the published version

Journal ref: Phys. Rev. E 101, 042125 (2020)

Decentralized Periodic Approach for Adaptive Fault Diagnosis in Distributed Systems

Authors: Latika Sarna, Sumedha Shenolikar, Poorva Kulkarni, Varsha Deshpande, Supriya Kelkar

Abstract: In this paper, Decentralized Periodic Approach for Adaptive Fault Diagnosis (DP-AFD) algorithm is proposed for fault diagnosis in distributed systems with arbitrary topology. Faulty nodes may be either unresponsive, may have either software or hardware faults. The proposed algorithm detects the faulty nodes situated in geographically distributed locations. This algorithm does not depend on a singl… ▽ More In this paper, Decentralized Periodic Approach for Adaptive Fault Diagnosis (DP-AFD) algorithm is proposed for fault diagnosis in distributed systems with arbitrary topology. Faulty nodes may be either unresponsive, may have either software or hardware faults. The proposed algorithm detects the faulty nodes situated in geographically distributed locations. This algorithm does not depend on a single node or leader to detect the faults in the system. However, it empowers more than one node to detect the fault-free and faulty nodes in the system. Thus, at the end of each test cycle, every fault-free node acts as a leader to diagnose faults in the system. This feature of the algorithm makes it applicable to any arbitrary network. After every test cycle of the algorithm, all the nodes have knowledge about faulty nodes and each node is tested only once. With this knowledge, there can be redistribution of load, which was earlier assigned to the faulty nodes. Also, the algorithm permits repaired node re-entry and new node entry. In a system of n nodes, the maximum number of faulty nodes can be (n-1) which is detected by DP-AFD algorithm. DP-AFD is periodic in nature which executes test cycles after regular intervals to detect the faulty nodes in the given distributed system. △ Less

Submitted 19 December, 2018; originally announced December 2018.

Comments: 19 pages, 13 figures, 1 table

ACM Class: C.2.4; C.2.5; C.2.6

Fault Diagnosis for Distributed Systems using Accuracy Technique

Authors: Poorva Kulkarni, Varsha Deshpande, Latika Sarna, Sumedha Shenolikar, Supriya Kelkar

Abstract: Distributed Systems involve two or more computer systems which may be situated at geographically distinct locations and are connected by a communication network. Due to failures in the communication link, faults arise which may make the entire system dysfunctional. To enable seamless operation of the distributed system, these faults need to be detected and located accurately. This paper examines v… ▽ More Distributed Systems involve two or more computer systems which may be situated at geographically distinct locations and are connected by a communication network. Due to failures in the communication link, faults arise which may make the entire system dysfunctional. To enable seamless operation of the distributed system, these faults need to be detected and located accurately. This paper examines various techniques of handling faults in distributed systems and proposes and innovative technique which uses percent accuracy for detecting faulty nodes in the system. Every node in the system acts as an initiator and votes for certifying faulty nodes in the system. This certification is done on the basis of percent accuracy value of each faulty node which should exceed a predefined threshold value to qualify node as faulty. As the threshold increases, the number of faulty nodes detected in the system reduces. This is a decentralized approach with no dependency on a single node to act as a leader for diagnosis. This technique is also applicable to ad-hoc networks, which are static in nature. △ Less

Submitted 19 December, 2018; originally announced December 2018.

Comments: 13 pages, 10 figures, 3 tables

ACM Class: C.2.4

Conformal Bootstrap Signatures of the Tricritical Ising Universality Class

Authors: Chethan N Gowdigere, Jagannath Santara, Sumedha

Abstract: We study the tricritical Ising universality class using conformal bootstrap techniques. By studying bootstrap constraints originating from multiple correlators on the CFT data of multiple OPEs, we are able to determine the scaling dimension of the spin field $Δ_σ$ in various non-integer dimensions $2 \le d \le 3$. $Δ_σ$ is connected to the critical exponent $η$ that governs the (tri-)critical beha… ▽ More We study the tricritical Ising universality class using conformal bootstrap techniques. By studying bootstrap constraints originating from multiple correlators on the CFT data of multiple OPEs, we are able to determine the scaling dimension of the spin field $Δ_σ$ in various non-integer dimensions $2 \le d \le 3$. $Δ_σ$ is connected to the critical exponent $η$ that governs the (tri-)critical behaviour of the two point function via the relation, $η= 2 - d + 2 Δ_σ$. Our results for $Δ_σ$ match with the exactly known values in two and three dimensions and are a conjecture for non-integer dimensions. We also compare our CFT results for $Δ_σ$ with $ε$-expansion results, available up to $ε^3$ order. Our techniques can be naturally extended to study higher-order multi-critical points. △ Less

Submitted 10 May, 2021; v1 submitted 28 November, 2018; originally announced November 2018.

Comments: version published in Physical Review D, improved and additional figures, expanded version

Journal ref: Phys. Rev. D 101, 116020 (2020)

A Stochastic model for dynamics of FtsZ filaments and the formation of Z-ring

Authors: Arabind Swain, A. V. Anil Kumar, Sumedha

Abstract: Understanding the mechanisms responsible for the formation and growth of FtsZ polymers and their subsequent formation of the $Z$-ring is important for gaining insight into the cell division in prokaryotic cells. In this work, we present a minimal stochastic model that qualitatively reproduces {\it in vitro} observations of polymerization, formation of dynamic contractile ring that is stable for a… ▽ More Understanding the mechanisms responsible for the formation and growth of FtsZ polymers and their subsequent formation of the $Z$-ring is important for gaining insight into the cell division in prokaryotic cells. In this work, we present a minimal stochastic model that qualitatively reproduces {\it in vitro} observations of polymerization, formation of dynamic contractile ring that is stable for a long time and depolymerization shown by FtsZ polymer filaments. In this stochastic model, we explore different mechanisms for ring breaking and hydrolysis. In addition to hydrolysis, which is known to regulate the dynamics of other tubulin polymers like microtubules, we find that the presence of the ring allows for an additional mechanism for regulating the dynamics of FtsZ polymers. Ring breaking dynamics in the presence of hydrolysis naturally induce rescue and catastrophe events in this model irrespective of the mechanism of hydrolysis. △ Less

Submitted 29 June, 2020; v1 submitted 8 October, 2018; originally announced October 2018.

Comments: Replaced with published version

Journal ref: Eur. Phys. J. E (2020) 43: 43

Unbounded Derivations in Bunce-Deddens-Toeplitz Algebras

Authors: Slawomir Klimek, Matt McBride, Sumedha Rathnayake, Kaoru Sakai, Honglin Wang

Abstract: In this paper we study decompositions and classification problems for unbounded derivations in Bunce-Deddens-Toeplitz and Bunce-Deddens algebras. We also look at implementations of these derivations on associated GNS Hilbert spaces. In this paper we study decompositions and classification problems for unbounded derivations in Bunce-Deddens-Toeplitz and Bunce-Deddens algebras. We also look at implementations of these derivations on associated GNS Hilbert spaces. △ Less

Submitted 26 July, 2018; originally announced July 2018.

Subject2Vec: Generative-Discriminative Approach from a Set of Image Patches to a Vector

Authors: Sumedha Singla, Mingming Gong, Siamak Ravanbakhsh, Frank Sciurba, Barnabas Poczos, Kayhan N. Batmanghelich

Abstract: We propose an attention-based method that aggregates local image features to a subject-level representation for predicting disease severity. In contrast to classical deep learning that requires a fixed dimensional input, our method operates on a set of image patches; hence it can accommodate variable length input image without image resizing. The model learns a clinically interpretable subject-lev… ▽ More We propose an attention-based method that aggregates local image features to a subject-level representation for predicting disease severity. In contrast to classical deep learning that requires a fixed dimensional input, our method operates on a set of image patches; hence it can accommodate variable length input image without image resizing. The model learns a clinically interpretable subject-level representation that is reflective of the disease severity. Our model consists of three mutually dependent modules which regulate each other: (1) a discriminative network that learns a fixed-length representation from local features and maps them to disease severity; (2) an attention mechanism that provides interpretability by focusing on the areas of the anatomy that contribute the most to the prediction task; and (3) a generative network that encourages the diversity of the local latent features. The generative term ensures that the attention weights are non-degenerate while maintaining the relevance of the local regions to the disease severity. We train our model end-to-end in the context of a large-scale lung CT study of Chronic Obstructive Pulmonary Disease (COPD). Our model gives state-of-the art performance in predicting clinical measures of severity for COPD. The distribution of the attention provides the regional relevance of lung tissue to the clinical measurements. △ Less

Submitted 28 June, 2018; originally announced June 2018.

Comments: MICCAI 2018

A Tight Approximation for Submodular Maximization with Mixed Packing and Covering Constraints

Authors: Eyal Mizrachi, Roy Schwartz, Joachim Spoerhase, Sumedha Uniyal

Abstract: Motivated by applications in machine learning, such as subset selection and data summarization, we consider the problem of maximizing a monotone submodular function subject to mixed packing and covering constraints. We present a tight approximation algorithm that for any constant $ε>0$ achieves a guarantee of $1-\frac{1}{\mathrm{e}}-ε$ while violating only the covering constraints by a multiplicat… ▽ More Motivated by applications in machine learning, such as subset selection and data summarization, we consider the problem of maximizing a monotone submodular function subject to mixed packing and covering constraints. We present a tight approximation algorithm that for any constant $ε>0$ achieves a guarantee of $1-\frac{1}{\mathrm{e}}-ε$ while violating only the covering constraints by a multiplicative factor of $1-ε$. Our algorithm is based on a novel enumeration method, which unlike previous known enumeration techniques, can handle both packing and covering constraints. We extend the above main result by additionally handling a matroid independence constraints as well as finding (approximate) pareto set optimal solutions when multiple submodular objectives are present. Finally, we propose a novel and purely combinatorial dynamic programming approach that can be applied to several special cases of the problem yielding not only {\em deterministic} but also considerably faster algorithms. For example, for the well studied special case of only packing constraints (Kulik {\em et. al.} [Math. Oper. Res. `13] and Chekuri {\em et. al.} [FOCS `10]), we are able to present the first deterministic non-trivial approximation algorithm. We believe our new combinatorial approach might be of independent interest. △ Less

Submitted 19 December, 2018; v1 submitted 29 April, 2018; originally announced April 2018.

A Tight Extremal Bound on the Lovász Cactus Number in Planar Graphs

Authors: Parinya Chalermsook, Andreas Schmid, Sumedha Uniyal

Abstract: A cactus graph is a graph in which any two cycles are edge-disjoint. We present a constructive proof of the fact that any plane graph $G$ contains a cactus subgraph $C$ where $C$ contains at least a $\frac{1}{6}$ fraction of the triangular faces of $G$. We also show that this ratio cannot be improved by showing a tight lower bound. Together with an algorithm for linear matroid parity, our bound im… ▽ More A cactus graph is a graph in which any two cycles are edge-disjoint. We present a constructive proof of the fact that any plane graph $G$ contains a cactus subgraph $C$ where $C$ contains at least a $\frac{1}{6}$ fraction of the triangular faces of $G$. We also show that this ratio cannot be improved by showing a tight lower bound. Together with an algorithm for linear matroid parity, our bound implies two approximation algorithms for computing "dense planar structures" inside any graph: (i) A $\frac{1}{6}$ approximation algorithm for, given any graph $G$, finding a planar subgraph with a maximum number of triangular faces; this improves upon the previous $\frac{1}{11}$-approximation; (ii) An alternate (and arguably more illustrative) proof of the $\frac{4}{9}$ approximation algorithm for finding a planar subgraph with a maximum number of edges. Our bound is obtained by analyzing a natural local search strategy and heavily exploiting the exchange arguments. Therefore, this suggests the power of local search in handling problems of this kind. △ Less

Submitted 15 March, 2019; v1 submitted 10 April, 2018; originally announced April 2018.

Comments: This result appeared in STACS19

Derivations and Spectral Triples on Quantum Domains II: Quantum Annulus

Authors: Slawomir Klimek, Matt McBride, Sumedha Rathnayake

Abstract: Continuing our study of spectral triples on quantum domains, we look at unbounded invariant and covariant derivations in the quantum annulus. In particular, we investigate whether such derivations can be implemented by operators with compact parametrices, a necessary condition in the definition of a spectral triple. Continuing our study of spectral triples on quantum domains, we look at unbounded invariant and covariant derivations in the quantum annulus. In particular, we investigate whether such derivations can be implemented by operators with compact parametrices, a necessary condition in the definition of a spectral triple. △ Less

Submitted 3 March, 2018; v1 submitted 17 October, 2017; originally announced October 2017.

Derivations and Spectral Triples on Quantum Domains I: Quantum Disk

Authors: Slawomir Klimek, Matt McBride, Sumedha Rathnayake, Kaoru Sakai, Honglin Wang

Abstract: We study unbounded invariant and covariant derivations on the quantum disk. In particular we answer the question whether such derivations come from operators with compact parametrices and thus can be used to define spectral triples. We study unbounded invariant and covariant derivations on the quantum disk. In particular we answer the question whether such derivations come from operators with compact parametrices and thus can be used to define spectral triples. △ Less

Submitted 24 September, 2017; v1 submitted 11 May, 2017; originally announced May 2017.

Journal ref: SIGMA 13 (2017), 075, 26 pages

Spectral Triples for nonarchimedean local fields

Authors: Slawomir Klimek, Sumedha Rathnayake, Kaoru Sakai

Abstract: Using associated trees, we construct a spectral triple for the C$^*$-algebra of continuous functions on the ring of integers $R$ of a nonarchimedean local field $F$ of characteristic zero, and investigate its properties. Remarkably, the spectrum of the spectral triple operator is closely related to the roots of a $q$-hypergeometric function. We also study a non compact version of this construction… ▽ More Using associated trees, we construct a spectral triple for the C$^*$-algebra of continuous functions on the ring of integers $R$ of a nonarchimedean local field $F$ of characteristic zero, and investigate its properties. Remarkably, the spectrum of the spectral triple operator is closely related to the roots of a $q$-hypergeometric function. We also study a non compact version of this construction for the C$^*$-algebra of continuous functions on $F$, vanishing at infinity. △ Less

Submitted 11 December, 2016; originally announced December 2016.

MSC Class: 46L87

Absence of first order transition in the random crystal field Blume-Capel model on a fully connected graph

Authors: Sumedha, Nabin K Jana

Abstract: In this paper we solve the Blume-Capel model on a complete graph in the presence of random crystal field with a distribution, $P(Δ_i) =pδ(Δ_i-Δ)+(1-p) δ(Δ_i+Δ)$, using large deviation techniques. We find that the first order transition of the pure system is destroyed for $0.046<p<0.954$ for all values of the crystal field, $Δ$. The system has a line of continuous transition for this range of $p$ f… ▽ More In this paper we solve the Blume-Capel model on a complete graph in the presence of random crystal field with a distribution, $P(Δ_i) =pδ(Δ_i-Δ)+(1-p) δ(Δ_i+Δ)$, using large deviation techniques. We find that the first order transition of the pure system is destroyed for $0.046<p<0.954$ for all values of the crystal field, $Δ$. The system has a line of continuous transition for this range of $p$ from $-\infty <Δ<\infty$. For values of $p$ outside this interval, the phase diagram of the system is similar to the pure model, with a tricritical point separating the line of first order and continuous transitions. We find that in this regime, the order vanishes for large $Δ$ for $p<0.046$(and for large $-Δ$ for $p>0.954$) even at zero temperature. △ Less

Submitted 23 November, 2016; v1 submitted 12 August, 2016; originally announced August 2016.

Comments: replaced with the accepted version; to be published in J. Phys. A

Journal ref: J. Phys. A: Math. Theor. 50 015003(2017)

An approximation algorithm for Uniform Capacitated k-Median problem with 1 + ε capacity violation

Authors: Jarosław Byrka, Bartosz Rybicki, Sumedha Uniyal

Abstract: We study the Capacitated k-Median problem, for which all the known constant factor approximation algorithms violate either the number of facilities or the capacities. While the standard LP-relaxation can only be used for algorithms violating one of the two by a factor of at least two, Shi Li [SODA'15, SODA'16] gave algorithms violating the number of facilities by a factor of 1+ε exploring properti… ▽ More We study the Capacitated k-Median problem, for which all the known constant factor approximation algorithms violate either the number of facilities or the capacities. While the standard LP-relaxation can only be used for algorithms violating one of the two by a factor of at least two, Shi Li [SODA'15, SODA'16] gave algorithms violating the number of facilities by a factor of 1+ε exploring properties of extended relaxations. In this paper we develop a constant factor approximation algorithm for Uniform Capacitated k-Median violating only the capacities by a factor of 1+ε. The algorithm is based on a configuration LP. Unlike in the algorithms violating the number of facilities, we cannot simply open extra few facilities at selected locations. Instead, our algorithm decides about the facility openings in a carefully designed dependent rounding process. △ Less

Submitted 23 November, 2015; originally announced November 2015.

Effect of random field disorder on the first order transition in $p$-spin interaction model

Authors: Sumedha, Sushant K. Singh

Abstract: We study the random field $p$-spin model with Ising spins on a fully connected graph using the theory of large deviations in this paper. This is a good model to study the effect of quenched random field on systems which have a sharp first order transition in the pure state. For $p=2$, the phase-diagram of the model, for bimodal distribution of the random field, has been well studied and is known t… ▽ More We study the random field $p$-spin model with Ising spins on a fully connected graph using the theory of large deviations in this paper. This is a good model to study the effect of quenched random field on systems which have a sharp first order transition in the pure state. For $p=2$, the phase-diagram of the model, for bimodal distribution of the random field, has been well studied and is known to undergo a continuous transition for lower values of the random field ($h$) and a first order transition beyond a threshold, $h_{tp}(\approx 0.439)$. We find the phase diagram of the model, for all $p \ge 2$, with bimodal random field distribution, using large deviation techniques. We also look at the fluctuations in the system by calculating the magnetic susceptibility. For $p=2$, beyond the tri-critical point in the regime of first order transition, we find that for $h_{tp}<h<0.447$, magnetic susceptibility increases rapidly (even though it never diverges) as one approaches the transition from the high temperature side. On the other hand, for $0.447<h \le 0.5$, the high temperature behaviour is well described by the Curie-Weiss law. For all $p \ge 2$, we find that for larger magnitudes of the random field ($h>h_o=1/p!$), the system does not show ferromagnetic order even at zero temperature. We find that the magnetic susceptibility for $p \ge 3$ is discontinuous at the transition point for $h<h_o$. △ Less

Submitted 9 September, 2015; v1 submitted 7 September, 2015; originally announced September 2015.

Comments: accepted for publication in Physica A; added extra references

A note on spectral properties of the $p$-adic tree

Authors: Slawomir Klimek, Sumedha Rathnayake, Kaoru Sakai

Abstract: We study the spectrum of the operator $D^*D$, where the operator $D$, introduced in \cite{KMR}, is a forward derivative on the $p$-adic tree, a weighted rooted tree associated to $\mathbb Z_p$ via Michon's correspondence. We show that the spectrum is closely related to the roots of a certain $q-$hypergeometric function and discuss the analytic continuation of the zeta function associated with… ▽ More We study the spectrum of the operator $D^*D$, where the operator $D$, introduced in \cite{KMR}, is a forward derivative on the $p$-adic tree, a weighted rooted tree associated to $\mathbb Z_p$ via Michon's correspondence. We show that the spectrum is closely related to the roots of a certain $q-$hypergeometric function and discuss the analytic continuation of the zeta function associated with $D^*D$. △ Less

Submitted 10 March, 2015; originally announced March 2015.

MSC Class: Primary 47A10; Secondary 47A11; 47A75; 46L89

An alternate view of complexity in k-SAT problems

Authors: Supriya Krishnamurthy, Sumedha

Abstract: The satisfiability threshold for constraint satisfaction problems is that value of the ratio of constraints (or clauses) to variables, above which the probability that a random instance of the problem has a solution is zero in the large system limit. Two different approaches to obtaining this threshold have been discussed in the literature - using first or second-moment methods which give rigorous… ▽ More The satisfiability threshold for constraint satisfaction problems is that value of the ratio of constraints (or clauses) to variables, above which the probability that a random instance of the problem has a solution is zero in the large system limit. Two different approaches to obtaining this threshold have been discussed in the literature - using first or second-moment methods which give rigorous bounds or using the non-rigorous but powerful replica-symmetry breaking (RSB) approach, which gives very accurate predictions on random graphs. In this paper, we lay out a different route to obtaining this threshold on a Bethe lattice. We need make no assumptions about the solution-space structure, a key assumption in the RSB approach. Despite this, our expressions and threshold values exactly match the best predictions of the cavity method under the 1-RSB assumption. Our method hence provides alternate interpretations as well as motivations for the key equations in the RSB approach. △ Less

Submitted 11 December, 2014; v1 submitted 8 December, 2014; originally announced December 2014.

Comments: 5 pages, 3 figures, typos corrected

Journal ref: Phys. Rev. E 92, 042144(2015)

The Quantum Pair of Pants

Authors: Slawomir Klimek, Matt McBride, Sumedha Rathnayake, Kaoru Sakai

Abstract: We compute the spectrum of the operator of multiplication by the complex coordinate in a Hilbert space of holomorphic functions on a disk with two circular holes. Additionally we determine the structure of the $C^*$-algebra generated by that operator. The algebra can be considered as the quantum pair of pants. We compute the spectrum of the operator of multiplication by the complex coordinate in a Hilbert space of holomorphic functions on a disk with two circular holes. Additionally we determine the structure of the $C^*$-algebra generated by that operator. The algebra can be considered as the quantum pair of pants. △ Less

Submitted 10 February, 2015; v1 submitted 2 October, 2014; originally announced October 2014.

Journal ref: SIGMA 11 (2015), 012, 22 pages

A $P$-Adic Spectral Triple

Authors: Slawomir Klimek, Matt McBride, Sumedha Rathnayake

Abstract: We construct a spectral triple for the C$^*$-algebra of continuous functions on the space of $p$-adic integers by using a rooted tree obtained from coarse-grained approximation of the space, and the forward derivative on the tree. Additionally, we verify that our spectral triple satisfies the properties of a compact spectral metric space, and we show that the metric on the space of $p$-adic intege… ▽ More We construct a spectral triple for the C$^*$-algebra of continuous functions on the space of $p$-adic integers by using a rooted tree obtained from coarse-grained approximation of the space, and the forward derivative on the tree. Additionally, we verify that our spectral triple satisfies the properties of a compact spectral metric space, and we show that the metric on the space of $p$-adic integers induced by the spectral triple is equivalent to the usual $p$-adic metric. △ Less

Submitted 31 March, 2014; v1 submitted 27 March, 2014; originally announced March 2014.

A Value Region Problem for Continued Fractions and Discrete Dirac Equations

Authors: Slawomir Klimek, Matt McBride, Sumedha Rathnayake, Kaoru Sakai

Abstract: Motivated by applications in noncommutative geometry we prove several value range estimates for even convergents and tails, and odd reverse sequences of Stieltjes type continued fractions with bounded ratio of consecutive elements, and show how those estimates control growth of solutions of a system of discrete Dirac equations. Motivated by applications in noncommutative geometry we prove several value range estimates for even convergents and tails, and odd reverse sequences of Stieltjes type continued fractions with bounded ratio of consecutive elements, and show how those estimates control growth of solutions of a system of discrete Dirac equations. △ Less

Submitted 3 March, 2018; v1 submitted 23 December, 2013; originally announced December 2013.

An Algorithm to Improve Performance over Multihop Wireless Mesh Network

Authors: Sumedha Chokhandre, Urmila Shrawankar

Abstract: Transmission Control Protocol (TCP) is the dominant reliable transport protocol utilized in the Internet. Improving the performance of TCP associated with the presence of multi-hop is one of the research challenges in wireless mesh networks. Wireless mesh networks have large round trip time variations and these variations are dependent on the number of hops. In wireless mesh network, when congesti… ▽ More Transmission Control Protocol (TCP) is the dominant reliable transport protocol utilized in the Internet. Improving the performance of TCP associated with the presence of multi-hop is one of the research challenges in wireless mesh networks. Wireless mesh networks have large round trip time variations and these variations are dependent on the number of hops. In wireless mesh network, when congestion loss and wireless loss are co-existed the number of packets dropped increases and will have adverse effects on TCP and its congestion control mechanism which leads to low throughput. Here we have designed a new TCP scheme for multi-hop wireless mesh networks, by modifying the sender side congestion control functionality of TCP NewReno, which is tuned towards improving the performance of TCP. The simulation results show that TCP SAC has higher performance than TCP NewReno, Reno, Sack and Vegas in multi-hop wireless mesh networks. △ Less

Submitted 10 May, 2013; originally announced May 2013.

Comments: Pages: 05 Figures: 07

Journal ref: Journal of Computing, Volume 3, Issue 5, May 2011, ISSN 2151-9617, pp 155,159

Balanced K-SAT and Biased random K-SAT on trees

Authors: Sumedha, Supriya Krishnamurthy, Sharmistha Sahoo

Abstract: We study and solve some variations of the random K-satisfiability problem - balanced K-SAT and biased random K-SAT - on a regular tree, using techniques we have developed earlier(arXiv:1110.2065). In both these problems, as well as variations of these that we have looked at, we find that the SAT-UNSAT transition obtained on the Bethe lattice matches the exact threshold for the same model on a rand… ▽ More We study and solve some variations of the random K-satisfiability problem - balanced K-SAT and biased random K-SAT - on a regular tree, using techniques we have developed earlier(arXiv:1110.2065). In both these problems, as well as variations of these that we have looked at, we find that the SAT-UNSAT transition obtained on the Bethe lattice matches the exact threshold for the same model on a random graph for K=2 and is very close to the numerical value obtained for K=3. For higher K it deviates from the numerical estimates of the solvability threshold on random graphs, but is very close to the dynamical 1-RSB threshold as obtained from the first non-trivial fixed point of the survey propagation algorithm. △ Less

Submitted 9 December, 2012; originally announced December 2012.

Comments: 22 pages, 7 figures

Journal ref: Phys. Rev. E 87, 042130 (2013)

On the behaviour of random K-SAT on trees

Authors: Supriya Krishnamurthy, Sumedha

Abstract: We consider the K-satisfiability problem on a regular d-ary rooted tree. For this model, we demonstrate how we can calculate in closed form, the moments of the total number of solutions as a function of d and K, where the average is over all realizations, for a fixed assignment of the surface variables. We find that different moments pick out different 'critical' values of d, below which they dive… ▽ More We consider the K-satisfiability problem on a regular d-ary rooted tree. For this model, we demonstrate how we can calculate in closed form, the moments of the total number of solutions as a function of d and K, where the average is over all realizations, for a fixed assignment of the surface variables. We find that different moments pick out different 'critical' values of d, below which they diverge as the total number of variables on the tree goes to infinity and above which they decay. We show that K-SAT on the random graph also behaves similarly. We also calculate exactly the fraction of instances that have solutions for all K. On the tree, this quantity decays to 0 (as the number of variables increases) for any d>1. However the recursion relations for this quantity have a non-trivial fixed-point solution which indicates the existence of a different transition in the interior of an infinite rooted tree. △ Less

Submitted 27 April, 2012; v1 submitted 10 October, 2011; originally announced October 2011.

Comments: 22 pages, 5 figures,accepted for publication in J. Stat. Mech

Journal ref: J. Stat. Mech. (2012) P05009

Prolonging assembly through dissociation:A self assembly paradigm in microtubules

Authors: Sumedha, Michael F Hagan, Bulbul Chakraborty

Abstract: We study a one-dimensional model of microtubule assembly/disassembly in which GTP bound to tubulins within the microtubule undergoes stochastic hydrolysis. In contrast to models that only consider a cap of GTP-bound tubulin, stochastic hydrolysis allows GTP-bound tubulin remnants to exist within the microtubule. We find that these buried GTP remnants enable an alternative mechanism of recovery fro… ▽ More We study a one-dimensional model of microtubule assembly/disassembly in which GTP bound to tubulins within the microtubule undergoes stochastic hydrolysis. In contrast to models that only consider a cap of GTP-bound tubulin, stochastic hydrolysis allows GTP-bound tubulin remnants to exist within the microtubule. We find that these buried GTP remnants enable an alternative mechanism of recovery from shrinkage, and enhances fluctuations of filament lengths. Under conditions for which this alternative mechanism dominates, an increasing depolymerization rate leads to a decrease in dissociation rate and thus a net increase in assembly. △ Less

Submitted 13 April, 2011; v1 submitted 8 August, 2009; originally announced August 2009.

Comments: accepted for publication in Physical Review E

Journal ref: Phys. Rev. E 83, 051904 (2011)

A thermodynamic model for agglomeration of DNA-loo** proteins

Authors: Sumedha, Martin Weigt

Abstract: In this paper, we propose a thermodynamic mechanism for the formation of transcriptional foci via the joint agglomeration of DNA-loo** proteins and protein-binding domains on DNA: The competition between the gain in protein-DNA binding free energy and the entropy loss due to DNA loo** is argued to result in an effective attraction between loops. A mean-field approximation can be described an… ▽ More In this paper, we propose a thermodynamic mechanism for the formation of transcriptional foci via the joint agglomeration of DNA-loo** proteins and protein-binding domains on DNA: The competition between the gain in protein-DNA binding free energy and the entropy loss due to DNA loo** is argued to result in an effective attraction between loops. A mean-field approximation can be described analytically via a map** to a restricted random-graph ensemble having local degree constraints and global constraints on the number of connected components. It shows the emergence of protein clusters containing a finite fraction of all loo** proteins. If the entropy loss due to a single DNA loop is high enough, this transition is found to be of first order. △ Less

Submitted 20 October, 2008; v1 submitted 9 January, 2008; originally announced January 2008.

Comments: 12 pages, 5 figures, to app. in JSTAT

Journal ref: J. Stat. Mech. (2008) P11005

Unsupervised and semi-supervised clustering by message passing: Soft-constraint affinity propagation

Authors: Michele Leone, Sumedha, Martin Weigt

Abstract: Soft-constraint affinity propagation (SCAP) is a new statistical-physics based clustering technique. First we give the derivation of a simplified version of the algorithm and discuss possibilities of time- and memory-efficient implementations. Later we give a detailed analysis of the performance of SCAP on artificial data, showing that the algorithm efficiently unveils clustered and hierarchical… ▽ More Soft-constraint affinity propagation (SCAP) is a new statistical-physics based clustering technique. First we give the derivation of a simplified version of the algorithm and discuss possibilities of time- and memory-efficient implementations. Later we give a detailed analysis of the performance of SCAP on artificial data, showing that the algorithm efficiently unveils clustered and hierarchical data structures. We generalize the algorithm to the problem of semi-supervised clustering, where data are already partially labeled, and clustering assigns labels to previously unlabeled points. SCAP uses both the geometrical organization of the data and the available labels assigned to few points in a computationally efficient way, as is shown on artificial and biological benchmark data. △ Less

Submitted 15 September, 2008; v1 submitted 7 December, 2007; originally announced December 2007.

Comments: 11 pages, 13 pdf figures, to app. in EPJB

Journal ref: Eur. Phys. J. B (2008), published online 8 Oct. 2008

Clustering by soft-constraint affinity propagation: Applications to gene-expression data

Authors: Michele Leone, Sumedha, Martin Weigt

Abstract: Motivation: Similarity-measure based clustering is a crucial problem appearing throughout scientific data analysis. Recently, a powerful new algorithm called Affinity Propagation (AP) based on message-passing techniques was proposed by Frey and Dueck \cite{Frey07}. In AP, each cluster is identified by a common exemplar all other data points of the same cluster refer to, and exemplars have to ref… ▽ More Motivation: Similarity-measure based clustering is a crucial problem appearing throughout scientific data analysis. Recently, a powerful new algorithm called Affinity Propagation (AP) based on message-passing techniques was proposed by Frey and Dueck \cite{Frey07}. In AP, each cluster is identified by a common exemplar all other data points of the same cluster refer to, and exemplars have to refer to themselves. Albeit its proved power, AP in its present form suffers from a number of drawbacks. The hard constraint of having exactly one exemplar per cluster restricts AP to classes of regularly shaped clusters, and leads to suboptimal performance, {\it e.g.}, in analyzing gene expression data. Results: This limitation can be overcome by relaxing the AP hard constraints. A new parameter controls the importance of the constraints compared to the aim of maximizing the overall similarity, and allows to interpolate between the simple case where each data point selects its closest neighbor as an exemplar and the original AP. The resulting soft-constraint affinity propagation (SCAP) becomes more informative, accurate and leads to more stable clustering. Even though a new {\it a priori} free-parameter is introduced, the overall dependence of the algorithm on external tuning is reduced, as robustness is increased and an optimal strategy for parameter selection emerges more naturally. SCAP is tested on biological benchmark data, including in particular microarray data related to various cancer types. We show that the algorithm efficiently unveils the hierarchical cluster structure present in the data sets. Further on, it allows to extract sparse gene expression signatures for each cluster. △ Less

Submitted 29 November, 2007; v1 submitted 18 May, 2007; originally announced May 2007.

Comments: 11 pages, supplementary material: http://isiosf.isi.it/~weigt/scap_supplement.pdf

Journal ref: Bioinformatics 23, 2708 (2007)

New structural variation in evolutionary searches of RNA neutral networks

Authors: Sumedha, Olivier C Martin, Andreas Wagner

Abstract: RNA secondary structure is an important computational model to understand how genetic variation maps into phenotypic (structural) variation. Evolutionary innovation in RNA structures is facilitated by neutral networks, large connected sets of RNA sequences that fold into the same structure. Our work extends and deepens previous studies on neutral networks. First, we show that even the 1-mutant n… ▽ More RNA secondary structure is an important computational model to understand how genetic variation maps into phenotypic (structural) variation. Evolutionary innovation in RNA structures is facilitated by neutral networks, large connected sets of RNA sequences that fold into the same structure. Our work extends and deepens previous studies on neutral networks. First, we show that even the 1-mutant neighborhood of a given sequence (genotype) G0 with structure (phenotype) P contains many structural variants that are not close to P. This holds for biological and generic RNA sequences alike. Second, we analyze the relation between new structures in the 1-neighborhoods of genotypes Gk that are only a moderate Hamming distance k away from G0, and the structure of G0 itself, both for biological and for generic RNA structures. Third, we analyze the relation between mutational robustness of a sequence and the distances of structural variants near this sequence. Our findings underscore the role of neutral networks in evolutionary innovation, and the role that high robustness can play in diminishing the potential for such innovation. △ Less

Submitted 15 December, 2006; originally announced December 2006.

Comments: to be published in Biosystems

Journal ref: Biosystems 90, 475-485(2007)