A5079075574- vaccines and immunoinformatics approaches
- Peptidase Inhibition and Analysis
- Software Testing and Debugging Techniques
- Immunotherapy and Immune Responses
- Monoclonal and Polyclonal Antibodies Research
- Glycosylation and Glycoproteins Research
- Software System Performance and Reliability
- Software Engineering Research
- Adversarial Robustness in Machine Learning
- Parallel Computing and Optimization Techniques
- Software Reliability and Analysis Research
- Explainable Artificial Intelligence (XAI)
- Formal Methods in Verification
- Advanced Malware Detection Techniques
- Advanced Software Engineering Methodologies
- Machine Learning and Data Classification
- Advanced Neural Network Applications
- Blockchain Technology Applications and Security
- Speech Recognition and Synthesis
- Model-Driven Software Engineering Techniques
- Radiation Effects in Electronics
- Computational Drug Discovery Methods
- Advanced Proteomics Techniques and Applications
- Topic Modeling
- Software Engineering Techniques and Practices
University of Edinburgh
2016-2025
Government Medical College
2024
University of Glasgow
2023
Halliburton (United States)
2023
University of Oxford
2012-2013
Laboratoire d'Informatique de Grenoble
2010-2011
Université Joseph Fourier
2011
Université Grenoble Alpes
2010
University of Minnesota
2006-2008
University of Minnesota System
2006-2008
In black-box testing, one is interested in creating a suite of tests from requirements that adequately exercise the behavior software system without regard to internal structure implementation. current practice, adequacy black box test suites inferred by examining coverage on an executable artifact, either source code or model.In this paper, we define structural metrics directly high-level formal requirements. These provide objective, implementation-independent measures how well exercises...
In avionics and other critical systems domains, adequacy of test suites is currently measured using the MC/DC metric on source code (or a model in model-based development). We believe that rigor highly sensitive to structure implementation can therefore be misleading as criterion. investigate this hypothesis by empirically studying effect program coverage.
Executing, verifying and enforcing credible transactions on permissionless blockchains is done using smart contracts. A key challenge with contracts ensuring their correctness security. To address this challenge, we present a fully automated technique, SolAnalyser, for vulnerability detection over Solidity that uses both static dynamic analysis. Analysis techniques in the literature rely analysis high rate of false positives or lack support vulnerabilities like out gas, unchecked send,...
Deep learning models for medical image classification tasks are becoming widely implemented in AI-assisted diagnostic tools, aiming to enhance accuracy, reduce clinician workloads, and improve patient outcomes. However, their vulnerability adversarial attacks poses significant risks safety. Current attack methodologies use general techniques such as model querying or pixel value perturbations generate examples designed fool a model. These approaches may not adequately address the unique...
Abstract ProteinMPNN is widely used in protein design workflows due to its ability identify amino acid sequences that fold into specific 3D structures. In our work, we adjust proteins for a given structure with reduced immune-visibility cytotoxic T lymphocytes recognize via the MHC-I pathway. To achieve this, developed novel framework integrates Direct Preference Optimization (DPO)—a tuning method originally designed large language models—with peptide presentation predictions. This approach...
Abstract Tumor antigens can emerge through multiple mechanisms, including translation of noncoding genomic regions. This noncanonical category tumor has recently gained attention; however, our understanding how they recur within and between cancer types is still in its infancy. Therefore, we developed a proteogenomic pipeline based on deep learning de novo mass spectrometry (MS) to enable the discovery MHC class I–associated peptides (ncMAP) from Considering that emergence also involve...
Abstract The ability to predict whether a peptide will get presented on Major Histocompatibility Complex (MHC) class I molecules has profound implications in designing vaccines. Numerous deep learning-based predictors for presentation MHC exist with high levels of accuracy. However, these are treated as black-box functions, providing little insight into their decision making. To build turst predictors, it is crucial understand the rationale behind decisions human-interpretable explanations....
Software design patterns are widely used in software engineering to enhance productivity and maintainability.However, recent empirical studies revealed the high energy overhead these patterns. Our vision is automatically detect transform during compilation for better efficiency without impacting existing coding practices. In this paper, we propose compiler transformations two patterns, Observer Decorator, perform an initial evaluation of their efficiency.
Test adequacy metrics defined over the structure of a program, such as Modified Condition and Decision Coverage (MC/DC), are used to assess testing efforts. However, MC/DC can be “cheated” by restructuring program make it easier achieve desired coverage. This is concerning, given importance in assessing test suites for critical systems domains. In this work, we have explored impact implementation on efficacy satisfying criterion using four real-world avionics systems. Our results demonstrate...
Solidity is an object-oriented and high-level language for writing smart contracts that are used to execute, verify enforce credible transactions on permissionless blockchains. In the last few years, analysis of has raised considerable interest numerous techniques have been proposed check presence vulnerabilities in them. Current lack traceability source code widely differing work flows. There no single unifying framework analysis, instrumentation, optimisation generation at level. this...
As product life-cycles become shorter and the scale complexity of systems increase, accelerating execution large test suites gains importance. Existing research has primarily focussed on techniques that reduce size suite. By contrast, we propose a technique accelerates execution, allowing to run in fraction original time, by parallel with Graphics Processing Unit (GPU).
In civil avionics, obtaining DO-178B certification for highly critical airborne software requires that the adequacy of code testing effort be measured using a structural coverage criterion known as Modified Condition and Decision Coverage (MC/DC). We hypothesized effectiveness MC/DC metric is sensitive to structure implementation can therefore problematic test criterion. tested this hypothesis by evaluating fault-finding ability MC/DC-adequate suites on five industrial systems (flight...
Model-based software development is gaining interest in domains such as avionics, space, and automotives.The model serves the central artifact for efforts (such as, code generation), therefore, it crucial that be extensively validated.Automatic generation of interaction test suites a candidate partial automation this validation task.Interaction testing combinatorial approach systematically tests all t-way combinations inputs system.In paper, we report how well (2-way through 5-way suites)...
Embedded software is found everywhere from our highly visible mobile devices to the confines of car in form smart sensors. companies are under huge pressure produce safe applications that limit risks, and testing absolutely critical alleviate concerns regarding safety user privacy. This requires using large test suites throughout development process, increasing time-to-market ultimately hindering competitivity.
Library migration is a challenging problem, where most existing approaches rely on prior knowledge. This can be, for example, information derived from changelogs or statistical models of API usage.
Explainable AI (XAI) techniques have been widely used to help explain and understand the output of deep learning models in fields such as image classification Natural Language Processing. Interest using XAI learning-based Automatic Speech Recognition (ASR) is emerging. But there not enough evidence on whether these explanations can be trusted. To address this, we adapt a state-of-the-art technique from domain, Local Interpretable Model-Agnostic Explanations (LIME), model trained for...
The rapidly advancing field of Explainable Artificial Intelligence (XAI) aims to tackle the issue trust regarding use complex black-box deep learning models in real-world applications. Existing post-hoc XAI techniques have recently been shown poor performance on medical data, producing unreliable explanations which are infeasible for clinical use. To address this, we propose an ante-hoc approach based concept bottleneck introduces first time concepts into classification pipeline, allowing...
Proteins have an arsenal of medical applications that include disrupting protein interactions, acting as potent vaccines, and replacing genetically deficient proteins. While therapeutics must avoid triggering unwanted immune-responses, vaccines should support a robust immune-reaction targeting broad range pathogen variants. Therefore, computational methods modifying proteins' immunogenicity without function are needed. many components the immune-system can be involved in reaction, we focus...
This paper proposes a fully explainable approach to speaker verification (SV), task that fundamentally relies on individual characteristics. The opaque use of attributes in current SV systems raises concerns trust. Addressing this, we propose an attribute-based system identifies speakers by comparing personal such as gender, nationality, and age extracted automatically from voice recordings. We believe this better aligns with human reasoning, making it more understandable than traditional...
ProteinMPNN is crucial in many protein design pipelines, identifying amino acid (AA) sequences that fold into given 3D backbone structures. We explore the context of designing therapeutic proteins need to avoid triggering unwanted immune reactions. More specifically, we focus on intra-cellular face challenge evading detection by Cytotoxic T-lymphocytes (CTLs) detect their presence via MHC Class I (MHC-I) pathway. To reduce visibility designed this immune-system component, develop a framework...