Total: 16
Named Entity Recognition models perform well on benchmark datasets but fail to generalize well even in the same domain. The goal of my th esis is to quantify the degree of in-domain generalization in NER, probe models for entity name vs. context learning and finally improve their robustness, focusing on the recognition of ethnically diverse entities and new entities over time when the models are deployed.
The doctoral work summarized here is an application of Artificial Intelligence (AI) for social good. The successful implementation would contribute towards low-cost, faster monitoring of road conditions across different nations, resulting in safer roads for everyone. Additionally, the study provides recommendations for re-using the road image data and the Deep Learning models released by any country for detecting road damage in other countries.
Generative models are used in a variety of applications that require diverse output. Yet, models are primarily optimised for sample fidelity and mode coverage. My work aims to increase the output diversity of generative models for multi-solution tasks. Previously, we analysed the use of generative models in artistic settings and how its objective diverges from distribution fitting. For specific use cases, we quantified the limitations of generative models. Future work will focus on adapting generative modelling for downstream tasks that require a diverse set of high-quality artefacts.
Contextual word representations generated by language models learn spurious associations present in the training corpora. Adversaries can exploit these associations to reverse-engineer the private attributes of entities mentioned in the training corpora. These findings have led to efforts towards minimizing the privacy risks of language models. However, existing approaches lack interpretability, compromise on data utility and fail to provide privacy guarantees. Thus, the goal of my doctoral research is to develop interpretable approaches towards privacy preservation of text representations that maximize data utility retention and guarantee privacy. To this end, I aim to study and develop methods to incorporate steganographic modifications within the vector geometry to obfuscate underlying spurious associations and retain the distributional semantic properties learnt during training.
Classical measurements and modelling that underpin present flood warning and alert systems are based on fixed and spatially restricted static sensor networks. Computationally expensive physics-based simulations are often used that can't react in real-time to changes in environmental conditions. We want to explore contemporary artificial intelligence (AI) for predicting flood risk in real time by using a diverse range of data sources. By combining heterogeneous data sources, we aim to nowcast rapidly changing flood conditions and gain a grater understanding of urgent humanitarian needs.
My work studies the problem of generating scenarios to evaluate interaction between humans and robots. I expect these interactions to grow in complexity as robots become more intelligent and enter our daily lives. However, evaluating such interactions only through user studies, which are the de facto evaluation method in human-robot interaction, will quickly become infeasible as the number of possible scenarios grows exponentially with scenario complexity. Therefore, I propose automatically generating scenarios in simulation to explore the diverse possibility space of scenarios to better understand interaction and avoid costly failures in real world settings.
Broadly speaking, I hope to dedicate my PhD to improving our understanding of algorithmic economics with the ultimate goal of building welfare improving decentralized technology for markets. In the following pages, I describe how my past work has built on the existing literature to get closer to the goal of creating such technologies, and describe what research paths this work opens up for the rest of my PhD. I believe that my research has the potential to provide algorithmic solutions to problems in machine learning, optimization, and game theory, and can be used to improve the efficiency of online marketplaces.
Recently, explanation methods have been proposed to evaluate the predictions of Graph Neural Networks on the task of link prediction. Evaluating explanation quality is difficult without ground truth explanations. This thesis is focused on providing a method, including datasets and scoring metrics, to quantitatively evaluate explanation methods on link prediction on Knowledge Graphs.
My dissertation investigates the computation of Bayes-Nash equilibria in auctions via multiagent learning. A particular focus lies on the game-theoretic analysis of learned gradient dynamics in such markets. This requires overcoming several technical challenges like non-differentiable utility functions and infinite-dimensional strategy spaces. Positive results may open the door for wide-ranging applications in Market Design and the economic sciences.
This study focuses on the Nesterov's accelerated quasi-Newton (NAQ) method in the context of deep neural networks (DNN) and its applications. The thesis objective is to confirm the robustness and efficiency of Nesterov's acceleration to quasi-Netwon (QN) methods by developing practical algorithms for different fields of optimization problems.
The entertainment industry, as well as the field of Computer Graphics, frequently faces the issue of creating large virtual crowds that would populate a scene. One of the ways to achieve that, particularly with modern rendering techniques, is by using simulation -- this, however, is nontrivial to design and control. The main goal of my PhD work is working towards the creation of a tool enabling the creation of virtual crowds that one can interact with, and we believe the best way to that is through Multiagent Reinforcement Learning techniques. These animated crowds can then be used both in movies and video games. Especially for the latter, it is highly desirable that both the crowd as a whole, as well as the individual characters, can react to the user's input in real time.
Artificial Intelligence has aimed to give the systems or agents, the ability to learn, perceive, recognize, plan, reason and act. Affective Computing has brought into focus the importance of giving AI systems, the capability to perceive, detect, utilize and generate emotion, affect, sentiment or feelings. To have a meaningful human-computer interaction, we need to design and develop a more socially intelligent and affective AI. My doctoral research goal is to delve deeper into some of these aspects, firstly by surveying computational models implemented in AI that uses emotion in decision-making or behaviour; secondly, by creating new model to predict social event context and affect in group videos; thirdly, to predict the social identities in visual scenes; and lastly to combine information about context, identities, behaviour and emotion in a social interaction scene to predict social incoherence and to recommend appropriate behaviour.
My PhD research focus is to produce a critical review of literature in Algorithmic Impact Assessment (AIA) and to develop an AIA tool that can be used to evaluate potential unintended impact of AI systems.
Tropical cyclones (TC) are extreme weather phenomena that bring heavy disasters to humans. Existing forecasting techniques contain computationally intensive dynamical models and statistical methods with complex inputs, both of which have bottlenecks in intensity forecasting, and we aim to create data-driven methods to break this forecasting bottleneck. The research goal of my PhD topic is to introduce novel methods to provide accurate and trustworthy forecasting of TC by developing interpretable machine learning models to analyze the characteristics of TC from multiple sources of data such as satellite remote sensing and observations.
My doctoral research focuses on understanding semantic knowledge in neural network models trained solely to predict natural language (referred to as language models, or LMs), by drawing on insights from the study of concepts and categories grounded in cognitive science. I propose a framework inspired by 'inductive reasoning,' a phenomenon that sheds light on how humans utilize background knowledge to make inductive leaps and generalize from new pieces of information about concepts and their properties. Drawing from experiments that study inductive reasoning, I propose to analyze semantic inductive generalization in LMs using phenomena observed in human-induction literature, investigate inductive behavior on tasks such as implicit reasoning and emergent feature recognition, and analyze and relate induction dynamics to the learned conceptual representation space.
Collaborative robots (i.e., "cobots") and machine learning-based virtual agents are increasingly entering the human workspace with the aim of increasing productivity, enhancing safety, and improving the quality of our lives. These agents will dynamically interact with a wide variety of people in dynamic and novel contexts, increasing the prevalence of human-machine teams in healthcare, manufacturing, and search-and-rescue. In this research, we enhance the mutual understanding within a human-machine team by enabling cobots to understand heterogeneous teammates via person-specific embeddings, identifying contexts in which xAI methods can help improve team mental model alignment, and enabling cobots to effectively communicate information that supports high-performance human-machine teaming.