-
Digital Evolution: Novo Nordisk's Shift to Ontology-Based Data Management
Authors:
Shawn Zheng Kai Tan,
Shounak Baksi,
Thomas Gade Bjerregaard,
Preethi Elangovan,
Thrishna Kuttikattu Gopalakrishnan,
Darko Hric,
Joffrey Joumaa,
Beidi Li,
Kashif Rabbani,
Santhosh Kannan Venkatesan,
Joshua Daniel Valdez,
Saritha Vettikunnel Kuriakose
Abstract:
Biomedical data is growing exponentially, and managing it is increasingly challenging. While Findable, Accessible, Interoperable and Reusable (FAIR) data principles provide guidance, their adoption has proven difficult, especially in larger enterprises like pharmaceutical companies. In this manuscript, we describe how we leverage an Ontology-Based Data Management (OBDM) strategy for digital transf…
▽ More
Biomedical data is growing exponentially, and managing it is increasingly challenging. While Findable, Accessible, Interoperable and Reusable (FAIR) data principles provide guidance, their adoption has proven difficult, especially in larger enterprises like pharmaceutical companies. In this manuscript, we describe how we leverage an Ontology-Based Data Management (OBDM) strategy for digital transformation in Novo Nordisk Research & Early Development. Here, we include both our technical blueprint and our approach for organizational change management. We further discuss how such an OBDM ecosystem plays a pivotal role in the organizations digital aspirations for data federation and discovery fuelled by artificial intelligence. Our aim for this paper is to share the lessons learned in order to foster dialogue with parties navigating similar waters while collectively advancing the efforts in the fields of data management, semantics and data driven drug discovery.
△ Less
Submitted 10 May, 2024; v1 submitted 8 May, 2024;
originally announced May 2024.
-
Social Evolution of Published Text and The Emergence of Artificial Intelligence Through Large Language Models and The Problem of Toxicity and Bias
Authors:
Arifa Khan,
P. Saravanan,
S. K Venkatesan
Abstract:
We provide a birds eye view of the rapid developments in AI and Deep Learning that has led to the path-breaking emergence of AI in Large Language Models. The aim of this study is to place all these developments in a pragmatic broader historical social perspective without any exaggerations while at the same time without any pessimism that created the AI winter in the 1970s to 1990s. We also at the…
▽ More
We provide a birds eye view of the rapid developments in AI and Deep Learning that has led to the path-breaking emergence of AI in Large Language Models. The aim of this study is to place all these developments in a pragmatic broader historical social perspective without any exaggerations while at the same time without any pessimism that created the AI winter in the 1970s to 1990s. We also at the same time point out toxicity, bias, memorization, sycophancy, logical inconsistencies, hallucinations that exist just as a warning to the overly optimistic. We note here that just as this emergence of AI seems to occur at a threshold point in the number of neural connections or weights, it has also been observed that human brain and especially the cortex region is nothing special or extraordinary but simply a case of scaled-up version of the primate brain and that even the human intelligence seems like an emergent phenomena of scale.
△ Less
Submitted 17 May, 2024; v1 submitted 11 February, 2024;
originally announced February 2024.
-
Building Flexible, Scalable, and Machine Learning-ready Multimodal Oncology Datasets
Authors:
Aakash Tripathi,
Asim Waqas,
Kavya Venkatesan,
Yasin Yilmaz,
Ghulam Rasool
Abstract:
The advancements in data acquisition, storage, and processing techniques have resulted in the rapid growth of heterogeneous medical data. Integrating radiological scans, histopathology images, and molecular information with clinical data is essential for develo** a holistic understanding of the disease and optimizing treatment. The need for integrating data from multiple sources is further prono…
▽ More
The advancements in data acquisition, storage, and processing techniques have resulted in the rapid growth of heterogeneous medical data. Integrating radiological scans, histopathology images, and molecular information with clinical data is essential for develo** a holistic understanding of the disease and optimizing treatment. The need for integrating data from multiple sources is further pronounced in complex diseases such as cancer for enabling precision medicine and personalized treatments. This work proposes Multimodal Integration of Oncology Data System (MINDS) - a flexible, scalable, and cost-effective metadata framework for efficiently fusing disparate data from public sources such as the Cancer Research Data Commons (CRDC) into an interconnected, patient-centric framework. MINDS offers an interface for exploring relationships across data types and building cohorts for develo** large-scale multimodal machine learning models. By harmonizing multimodal data, MINDS aims to potentially empower researchers with greater analytical ability to uncover diagnostic and prognostic insights and enable evidence-based personalized care. MINDS tracks granular end-to-end data provenance, ensuring reproducibility and transparency. The cloud-native architecture of MINDS can handle exponential data growth in a secure, cost-optimized manner while ensuring substantial storage optimization, replication avoidance, and dynamic access capabilities. Auto-scaling, access controls, and other mechanisms guarantee pipelines' scalability and security. MINDS overcomes the limitations of existing biomedical data silos via an interoperable metadata-driven approach that represents a pivotal step toward the future of oncology data integration.
△ Less
Submitted 22 December, 2023; v1 submitted 30 September, 2023;
originally announced October 2023.
-
ConStaBL -- A Fresh Look at Software Engineering with State Machines
Authors:
Karthika Venkatesan,
Sujit Kumar Chakrabarti
Abstract:
Statechart is a visual modelling language for systems. In this paper, we extend our earlier work on modular statecharts with local variables and present an updated operational semantics for statecharts with concurrency. Our variant of the statechart has local variables, which interact significantly with the remainder of the language semantics. Our semantics does not allow transition conflicts in s…
▽ More
Statechart is a visual modelling language for systems. In this paper, we extend our earlier work on modular statecharts with local variables and present an updated operational semantics for statecharts with concurrency. Our variant of the statechart has local variables, which interact significantly with the remainder of the language semantics. Our semantics does not allow transition conflicts in simulations and is stricter than most other available semantics of statecharts in that sense. It allows arbitrary interleaving of concurrently executing action code, which allows more precise modelling of systems and upstream analysis of the same. We present the operational semantics in the form of the simulation algorithm. We also establish the criteria based on our semantics for defining conflicting transitions and valid simulations. Our semantics is executable and can be used to simulate statechart models and verify their correctness. We present a preliminary setup to carry out fuzz testing of Statechart models, an idea that does not seem to have a precedent in literature. We have used our simulator in conjunction with a well-known fuzzer to do fuzz testing of statechart models of non-trivial sizes and have found issues in them that would have been hard to find through inspection.
△ Less
Submitted 11 July, 2023; v1 submitted 7 July, 2023;
originally announced July 2023.
-
Experimental Study of Forced Synchronization and Cross-coupling in a Liquid-Fuelled Gas Turbine Combustor at Elevated Pressure
Authors:
Mitchell L. Passarelli,
Askar Kazbekov,
Victor Salazar,
Krishna Venkatesan,
Adam M. Steinberg
Abstract:
The effects of external forcing on a turbulent, liquid-fuelled, swirl-stabilized gas turbine combustor operating at a pressure of approximately 1 MPa are explored experimentally. In particular, the dynamics and coupling between the hydrodynamics, heat release rate and acoustics are compared for various forcing amplitudes at a fixed forcing frequency $f_f$. The hydrodynamics were characterized via…
▽ More
The effects of external forcing on a turbulent, liquid-fuelled, swirl-stabilized gas turbine combustor operating at a pressure of approximately 1 MPa are explored experimentally. In particular, the dynamics and coupling between the hydrodynamics, heat release rate and acoustics are compared for various forcing amplitudes at a fixed forcing frequency $f_f$. The hydrodynamics were characterized via laser Mie scattering from droplets in the fuel spray, while the heat release rate was qualitatively measured using chemiluminescence (CL) emissions in the 312 $\pm$ 12.5 nm wavelength range, both at 10 kHz. The dynamics at the frequencies of interest were extracted using spectral proper orthogonal decomposition (SPOD). In the unforced case, the spray and CL oscillations exhibited similar dynamics, dominated by oscillations at frequency $f_0$, whereas the pressure fluctuations were predominantly at $f_P$. As the forcing amplitude was increased from zero, the spray and CL exhibited changes in their power spectra characteristic of the suppression route to synchronization. The pressure fluctuations, however, were observed to follow the phase-locking route to synchronization. In contrast with expectations from synchronization theory, the amplitude of the pressure fluctuations increased significantly not only after lock-on, but also as the frequency detuning with $f_f$ decreased. It is shown that this increase in amplitude is not due to intermittency in the frequency of the pressure oscillations. The simultaneous occurrence of phase-locking and suppression illustrates the rich variety of dynamics that can occur in practical combustor systems. In addition, the amplification of the pressure oscillations based on the frequency detuning with the forcing suggests that classical reasoning based on the Rayleigh Index may not be sufficient to understand the high amplitude behaviour of multimodal systems.
△ Less
Submitted 1 June, 2022;
originally announced June 2022.
-
Cross-Frequency Coupling During Thermoacoustic Oscillations in a Pressurized Aeronautical Gas Turbine Model Combustor
Authors:
Mitchell L. Passarelli,
Timothy M. Wabel,
Arin Cross,
Krishna Venkatesan,
Adam M. Steinberg
Abstract:
This paper demonstrates cross-frequency coupling between pressure, heat release rate, fuel spray and velocity oscillations in a model aeronautical gas turbine combustor operating at a pressure of approximately 10 atm. Heat release rate was characterized by 10 kHz chemiluminescence (CL) imaging of several species. Stereoscopic particle image velocimetry and laser Mie scattering from the fuel drople…
▽ More
This paper demonstrates cross-frequency coupling between pressure, heat release rate, fuel spray and velocity oscillations in a model aeronautical gas turbine combustor operating at a pressure of approximately 10 atm. Heat release rate was characterized by 10 kHz chemiluminescence (CL) imaging of several species. Stereoscopic particle image velocimetry and laser Mie scattering from the fuel droplets were used to measure the gas velocity and spray dynamics, respectively, at 5 kHz. The pressure fluctuations were dominated by oscillations at a frequency $f_0$, whereas the spray, CL and velocity oscillated at approximately $2f_0$. All of these oscillations were nonstationary, exhibiting changes in frequency and amplitude. Comparing the time evolution of the dominant frequencies and amplitudes indicates a behavior consistent with mutually coupled self-oscillators; the observed dynamics of the 1:2 super-harmonic coupling is consistent synchronization via oscillation death. Furthermore, increases in the frequency of the ca. $f_0$ velocity oscillations away from the harmonic ratio (increased frequency detuning) were correlated with decreases in the power of the $f_0$ pressure oscillations. The corresponding nonreacting flow had a natural hydrodynamic mode at a frequency slightly greater than $2f_0$. Hence, the data are consistent with the $f_0$ acoustic mode "pulling" the hydrodynamic frequency towards the super-harmonic ratio.
△ Less
Submitted 17 June, 2020;
originally announced June 2020.
-
StaBL - State Based Language for Specification of Web Applications
Authors:
Karthika Venkatesan,
Sujit Kumar Chakrabarti
Abstract:
Context and motivation: Usage of Formal Specification languages is scarce in web application development as compared to safety critical/hardware systems. Question/problem: An apt formal specification language should provide the following features: Firstly, it should have well-defined semantics, so that specifications written in it can not be inherently ambiguous. Secondly, it should have tool supp…
▽ More
Context and motivation: Usage of Formal Specification languages is scarce in web application development as compared to safety critical/hardware systems. Question/problem: An apt formal specification language should provide the following features: Firstly, it should have well-defined semantics, so that specifications written in it can not be inherently ambiguous. Secondly, it should have tool support for automatic detection of specification bugs. Additionally, for domains like web development, it is important that specification formalisms build over familiar notations, as the benefits of learning highly mathematical notations in such domains are perceived to be low. Principal ideas/results: This work presents a State Based Language inspired by Statecharts called StaBL for specification of web applications, and how StaBL can be used for writing such specifications. We also present modifications to the language w.r.t Statechart which facilitate writing modular and scalable specification. Contribution: In particular, we present the feature of locally scoped variables with inter-state data-flow. We summarise our experience of develo** specifications with StaBL, which shows that StaBL specifications, on the one hand, are able to capture most essential elements of the functional aspects of a web application while foregoing much of the verbosity of a regular programming language.
△ Less
Submitted 8 January, 2019;
originally announced January 2019.
-
Building a Structured Query Engine
Authors:
Amanpreet Singh,
Karthik Venkatesan,
Simranjyot Singh Gill
Abstract:
Finding patterns in data and being able to retrieve information from those patterns is an important task in Information retrieval. Complex search requirements which are not fulfilled by simple string matching and require exploring certain patterns in data demand a better query engine that can support searching via structured queries. In this article, we built a structured query engine which suppor…
▽ More
Finding patterns in data and being able to retrieve information from those patterns is an important task in Information retrieval. Complex search requirements which are not fulfilled by simple string matching and require exploring certain patterns in data demand a better query engine that can support searching via structured queries. In this article, we built a structured query engine which supports searching data through structured queries on the lines of ElasticSearch. We will show how we achieved real time indexing and retrieving of data through a RESTful API and how complex queries can be created and processed using efficient data structures we created for storing the data in structured way. Finally, we will conclude with an example of movie recommendation system built on top of this query engine.
△ Less
Submitted 1 October, 2017;
originally announced October 2017.
-
Charge Transport and Conductance Switching of Redox-active Azulene Derivatives
Authors:
Florian Schwarz,
Michael Koch,
Georg Kastlunger,
Heinz Berke,
Robert Stadler,
Koushik Venkatesan,
Emanuel Loertscher
Abstract:
Azulene (Az) is a non-alternating, aromatic hydrocarbon composed of a five-membered, electron-rich and a seven-membered, electron-poor ring; an electron distribution that provides intrinsic redox activity. By varying the attachment points of the two electrode-bridging substituents to the Az centre, the influence of the redox functionality on charge transport is evaluated. The conductance of the 1,…
▽ More
Azulene (Az) is a non-alternating, aromatic hydrocarbon composed of a five-membered, electron-rich and a seven-membered, electron-poor ring; an electron distribution that provides intrinsic redox activity. By varying the attachment points of the two electrode-bridging substituents to the Az centre, the influence of the redox functionality on charge transport is evaluated. The conductance of the 1,3 Az derivative is at least one order of magnitude lower than those of the 2,6 Az and 4,7 Az derivatives, in agreement with density functional theory (DFT) calculations. In addition, only 1,3 Az exhibits pronounced nonlinear current-voltage characteristics with hysteresis, indicating a bias-dependent conductance switching. DFT identifies the LUMO to be nearest to the Fermi energy of the electrodes, but to be an active transport channel only in the case of the 2,6 and the 4,7 Az derivatives, whereas the 1,3 Az derivative uses the HOMO at low and the LUMO+1 at high bias. In return, the localized, weakly coupled LUMO of 1,3 Az creates a slow electron-hop** channel responsible for the voltage-induced switching due to the occupation of a single MO.
△ Less
Submitted 1 September, 2017;
originally announced September 2017.
-
Field-induced Conductance Switching by Charge-state Alternation in Organometallic Single-Molecule Junctions
Authors:
Florian Schwarz,
Georg Kastlunger,
Franziska Lissel,
Carolina Egler-Lucas,
Sergey N. Semenov,
Koushik Venkatesan,
Heinz Berke,
Robert Stadler,
Emanuel Loertscher
Abstract:
Charge transport through single molecules can be influenced by the charge and spin states of redox-active metal centres placed in the transport pathway. These molecular intrinsic properties are usually addressed by varying the molecules electrochemical and magnetic environment, a procedure that requires complex setups with multiple terminals. Here we show that oxidation and reduction of organometa…
▽ More
Charge transport through single molecules can be influenced by the charge and spin states of redox-active metal centres placed in the transport pathway. These molecular intrinsic properties are usually addressed by varying the molecules electrochemical and magnetic environment, a procedure that requires complex setups with multiple terminals. Here we show that oxidation and reduction of organometallic compounds containing either Fe, Ru or Mo centres can solely be triggered by the electric field applied to a two-terminal molecular junction. Whereas all compounds exhibit bias-dependent hysteresis, the Mo-containing compound additionally shows an abrupt voltage-induced conductance switching, yielding high to low current ratios exceeding 1000 at voltage stimuli of less than 1.0 V. DFT calculations identify a localized, redox active molecular orbital that is weakly coupled to the electrodes and closely aligned with the Fermi energy of the leads because of the spin-polarised ground state unique to the Mo centre. This situation opens an additional slow and incoherent hop** channel for transport, triggering a transient charging effect of the entire molecule and a strong hysteresis with unprecedented high low-to-high current ratios.
△ Less
Submitted 5 September, 2016;
originally announced September 2016.
-
High-Conductive Organometallic Molecular Wires with Delocalized Electron Systems Strongly Coupled to Metal Electrodes
Authors:
Florian Schwarz,
Georg Kastlunger,
Franziska Lissel,
Heike Riel,
Koushik Venkatesan,
Heinz Berke,
Robert Stadler,
Emanuel Lörtscher
Abstract:
Besides active, functional molecular building blocks such as diodes or switches, passive components as, e.g., molecular wires, are required to realize molecular-scale electronics. Incorporating metal centers in the molecular backbone enables the molecular energy levels to be tuned in respect to the Fermi energy of the electrodes. Furthermore, by using more than one metal center and sp-bridging lig…
▽ More
Besides active, functional molecular building blocks such as diodes or switches, passive components as, e.g., molecular wires, are required to realize molecular-scale electronics. Incorporating metal centers in the molecular backbone enables the molecular energy levels to be tuned in respect to the Fermi energy of the electrodes. Furthermore, by using more than one metal center and sp-bridging ligands, a strongly delocalized electron system is formed between these metallic "dopants", facilitating transport along the molecular backbone. Here, we study the influence of molecule--metal coupling on charge transport of dinuclear X(PP)$_2$FeC$_4$Fe(PP)$_2$X molecular wires (PP = Et$_2$PCH$_2$CH$_2$PEt$_2$); X = CN (1), NCS (2), NCSe (3), C$_4$SnMe$_3$ (4) and C$_2$SnMe$_3$ (5)) under ultra-high vacuum and variable temperature conditions. In contrast to 1 which showed unstable junctions at very low conductance ($8.1\cdot10^{-7}$ G$_0$), 4 formed a Au-C$_4$FeC$_4$FeC$_4$-Au junction 4' after SnMe$_3$ extrusion which revealed a conductance of $8.9\cdot10^{-3}$ G$_0$, three orders of magnitude higher than for 2 ($7.9\cdot10^{-6}$ G$_0$) and two orders of magnitude higher than for 3 ($3.8\cdot10^{-4}$ G$_0$). Density functional theory (DFT) confirmed the experimental trend in the conductance for the various anchoring motifs. The strong hybridization of molecular and metal states found in the C--Au coupling case enables the delocalized electronic system of the organometallic Fe$_2$ backbone to be extended over the molecule-metal interfaces to the metal electrodes to establish high-conductive molecular wires.
△ Less
Submitted 5 November, 2015;
originally announced November 2015.
