DaSCI Seminars
Are talks by an outstanding invited researcher who presents recent disruptive advances on AI. The seminars will last about 1 hour and 30 minutes (45min. Speaker + 30 min. for questions)
DaSCI Seminars 2024
Causethical ML: from theory to practice
Date: 19/11/2024
Abstract: In this talk I will give an overview of the role of causality in ethical machine learning, and in particular, in fair and explainable ML. In particular, I will first detail how to use causal reasoning to study fairness and interpretability problems in algorithmic decision making, stressing the main limitations that we encounter when aiming to address these problems in practice. Then, I will provide some hints about how to solve some of these practical limitations by using causal generative models. A novel class of deep generative models that do not only accurately fit observational data but can also provide accurate estimates to interventional and counterfactual queries. I will finally discuss the open challenges of designing such causal generative models.
Speaker: Isabel Valera is Full Professor of Machine Learning at the Department of Computer Science at Saarland University (Saarbrücken, Germany), and Adjunct Faculty at the MPI for Software Systems in Saarbrücken (Saarbrücken, Germany). She is the recipient of an ERC Starting Grant on “Society-Aware ML”, and a Fellow of the European Laboratory for Learning and Intelligent Systems (ELLIS). Previously, she was an independent group leader at the MPI for Intelligent Systems in Tübingen, Germany. She received her Ph.D. in 2014 and her MSc in 2012 from the University Carlos III in Madrid, Spain, and worked as a postdoctoral researcher at the MPI for Software Systems (Germany) and the University of Cambridge (UK). Her research focuses on the development of trustworthy machine learning methods that can be used in the real world. Her research can be broadly categorized into three main themes: fair, interpretable, and robust machine learning. Her research interests cover a wide range of ML approaches, including deep learning, probabilistic modeling, causal inference, time series analysis, and many more.
Recording: Causethical ML: from theory to practice
Distributed learning, semantics of information, generative models, multi-agent learning under communications constraints
Date: 30/10/2024
Abstract:
Speaker: Dr. Marios Kountouris is a leading researcher and academic in the area of telecommunications and communication networks. He is known for his contributions to information theory, wireless networks and the design of communication algorithms. He has worked at prestigious institutions such as the Laboratoire d’Informatique de Paris (LIX) at the École Polytechnique, France, where he has led research on 5G communication and networks, wireless networks and the optimisation of networks to improve their efficiency and capacity. His work covers both theoretical and applied aspects of telecommunications, and he has contributed significantly to areas such as multi-user networks, communication in low-power devices and mass communication systems. He has also been a prolific author for renowned conferences and journals and is frequently invited to speak at international events. Dr. Marios Kountouris is one of the newest members of the Andalusian Data Science and Computational Intelligence Research Institute (DaSCI) at the University of Granada and has been awarded one of the prestigious European Research Council (ERC) grants. His research at DaSCI is focused on advanced topics in telecommunications and artificial intelligence, reinforcing the leadership of this institute in cutting-edge technological areas.
This conference is part of the Project ‘Ethical, Responsible and General Purpose Artificial Intelligence: Applications in Risk Scenarios’ (IAFER) Exp.: TSI-100927-2023-1 funded through the creation of university-industry chairs (Enia Chairs), aimed at research and development of artificial intelligence, for its dissemination and training in the framework of the European Recovery, Transformation and Resilience Plan, funded by the European Union-Next Generation EU.
Research and Industrialisation of Augmented Reality Systems for Soft Organs in Abdominal MIS
Date: 08/10/2024
Abstract: Mini-invasive surgery (MIS) is done using a camera called endoscope, inserted along with surgical instruments through tiny incisions in the abdominal wall. The surgeon looks at a screen and handles the instruments directly or via a telemanipulator in robot-assisted MIS. Despite its numerous advantages, MIS presents issues as the structures internal to the target organs may be difficult to localise for the surgeon. These can be structures to target such as tumours or structures to spare such as vessels. We have proposed to aid localisation of hidden structures in MIS using augmented reality, which is done by transferring information available from the diagnostic imaging such as CT and MRI acquired before surgery. Transferring these information presents numerous scientific, technical and clinical challenges. This presentation describes the academic development of a possible solution and its transfer to the industry.
Speaker: Adrien Bartoli has been a Professor of Computer Science at Université Clermont Auvergne since fall 2009 and a member of Institut Universitaire de France since 2016. He is currently on leave as research scientist at the University Hospital of Clermont-Ferrand and as Chief Scientific Officer at SurgAR (2021-2024). He founded and leads the EnCoV (Endoscopy and Computer Vision) research group jointly with Michel Canis. He held an ERC Consolidator grant (2013-2018) and an ERC Proof-of-Concept grant (2018-2019). Previously, he was a CNRS research scientist at Institut Pascal since fall 2004 where he led ComSee, the Computer Vision research group, jointly with Thierry Chateau. He was a Visiting Professor in DIKU at the University of Copenhagen between 2006-2009 and a postdoctoral researcher in the Visual Geometry Group at the University of Oxford under Andrew Zisserman in 2004. Adrien Bartoli obtained his Habilitation Degree (HDR) from Université Blaise Pascal in June 2008. He completed his PhD in the Perception group at Inria Grenoble under Peter Sturm and Radu Horaud. Adrien Bartoli has received several awards including the 2004 Grenoble-INP PhD thesis prize, the 2008 CNRS médaille de bronze and the 2016 research prize from Université d’Auvergne. Adrien Bartoli’s main research interests include image registration and Shape-from-X for rigid and non-rigid scenarios, and machine learning within the field of theoretical and medical Computer Vision. He has published approximately 100 scientific papers and has been on the program committees for top-ranking conferences in the field. He is on the editorial board of IJCV and JAIR and was on the editorial board of IET CV and ELCVIA.
Recording: Research and Industrialisation of Augmented Reality Systems for Soft Organs in Abdominal MIS
AI for Democracy: Ethical and Political Challenges
Date: 11/09/2024
Abstract: After the recent publication of his new book “Why AI undermines democracy and what to do about it”, Mark Coeckelbergh proposes a discussion on «AI for Democracy: Ethical and Political Challenges»
Speaker: Dr. Mark Coeckelbergh
Recording: AI for Democracy: Ethical and Political Challenges
The role of artificial intelligence in achieving the Sustainable Development Goals
Date: 28/05/2024
Abstract: The emergence of artificial intelligence (AI) and its progressively wider impact on many sectors requires an assessment of its effect on the achievement of the Sustainable Development Goals. Using a consensus-based expert elicitation process, we find that AI can enable the accomplishment of 134 targets across all the goals, but it may also inhibit 59 targets. However, current research foci overlook important aspects. The fast development of AI needs to be supported by the necessary regulatory insight and oversight for AI-based technologies to enable sustainable development. Failure to do so could result in gaps in transparency, safety, and ethical standards.
Speaker: Dr. Ricardo Vinuesa is an Associate Professor at the Department of Engineering Mechanics, KTH Royal Institute of Technology in Stockholm. He is also Vice Director of the KTH Digitalization Platform and Lead Faculty at the KTH Climate Action Centre. He studied Mechanical Engineering at the Polytechnic University of Valencia (Spain), and he received his PhD in Mechanical and Aerospace Engineering from the Illinois Institute of Technology in Chicago. His research combines numerical simulations and data-driven methods to understand, control and predict complex wall-bounded turbulent flows, such as the boundary layers developing around wings and urban environments. Dr. Vinuesa has received, among others, an ERC Consolidator Grant (the most prestigious research program in Europe), the TSFP Kasagi Award, the Goran Gustafsson Award for Young Researchers, the IIT Outstanding Young Alumnus Award, the SARES Young Researcher Award and he leads several large Horizon Europe projects. He is also a member of the Young Academy of Science of Spain.
Recording: The role of artificial intelligence in achieving the Sustainable Development Goals
Prevalidated ridge regression as a highly-efficient drop-in replacement for logistic regression for high-dimensional data
Date: 07/05/2024
Abstract: Linear models are widely used in classification, and are particularly effective for high-dimensional data where linear decision boundaries / separating hyperplanes are often effective for separating classes, even for complex data. A recent example of a technique effectively utilising linear classifiers is the ROCKET family of classifiers for time series classification. One reason that the ROCKET family is so fast is due to its use of a linear classifier based around standard squared-error ridge regression. Fitting a linear model based on squared-error is significantly faster and more stable than fitting a standard regularised multinomial logistic regression based on logarithmic-loss (i.e., regularised maximum likelihood), as in the latter case the solutions can only be found via a numerical search. While fast, one drawback of using squared-error ridge-regression is that it is unable to produce probabilistic predictions. I will demonstrate some very recent work on how to use regular ridge-regression to train L2-regularized multinomial logistic regression models for very large numbers of features, including choosing a suitable degree of regularization, with a time complexity that is no greater than single ordinary least-squares fit. This in contrast to logistic regression, which requires a full refit for every value of regularisation parameter considered, and every fold used for cross-validation. Using our new approach allows for models based on linear classifier technology to provide well calibrated probabilistic predictions with minimal additional computational overhead. If time permits, I will also discuss some thoughts on when such linear classifiers would be expected to perform well.
Speaker: Daniel Schmidt is an Associate Professor of Computer Science at the Department of Data Science and AI, Monash University, Australia. He obtained his PhD in the area of information theoretic statistical inference in 2008 from Monash University. From 2009 to 2018 he was employed at the University of Melbourne, working in the field of statistical genomics (GWAS, epigenetics and cancer). Since 2018 he has been employed at Monash University in a teaching and research position. His research interests are primarily time series classification and forecasting, particularly at scale, and Bayesian inference and MCMC, with an emphasis on sparsity and shrinkage, Bayesian optimisation and Bayesian function approximation. He also has a keen interest in the best ways to provide statistical/machine learning education.
Green AI
Date: 09/04/2024
Abstract: The computations required for deep learning research have been doubling every few months, resulting in an estimated 5,000x increase from 2018 to 2022. This trend has led to unprecedented success in a range of AI tasks. In this talk I will discuss a few troubling side-effects of this trend, touching on issues of lack of inclusiveness within the research community, and an increasingly large environmental footprint. I will then present Green AI – an alternative approach to help mitigate these concerns. Green AI is composed of two main ideas: increased reporting of computational budgets, and making efficiency an evaluation criterion for research alongside accuracy and related measures. I will focus on the latter topic, discussing various recent approaches for reducing the computational cost of AI, some based on reducing LLM cache size, others on showing that smaller LMs can be as accurate as larger ones. This is joint work with Michael Hassid, Yossi Adi, Matanel Oren, Tal Remez, Jesse Dodge, Noah A. Smith, Oren Etzioni and Jonas Gehring.
Speaker: Roy Schwartz is a senior lecturer (assistant professor) at the School of Computer Science and Engineering at The Hebrew University of Jerusalem (HUJI). Roy studies natural language processing and artificial intelligence. Prior to joining HUJI, Roy was a postdoc (2016-2019) and then a research scientist (2019-2020) at the Allen institute for AI and at The University of Washington, where he worked with Noah A. Smith. Roy completed his Ph.D. in 2016 at HUJI, where he worked with Ari Rappoport. Roy’s work has appeared on the cover of the CACM Magazine, and has been featured, among others, in New York Times , MIT Tech Review and Forbes.
Recording: Green AI
Reverse engineering of the human brain mechanisms of object perception and learning
Date: 06/02/2024
Abstract: The human species is engaged in a major scientific quest: to understand the neural mechanisms of human (primate) intelligence. Recent advances in multiple subfields of brain research suggest that the next key steps in this quest will be the construction of real-world, systemic-level network models that aim to abstract, emulate, and explain the primate neural mechanisms underlying natural intelligent behavior. In this talk, I will describe the history of how neuroscience, cognitive science, and computer science converged to create image-specific, computationally computable, deep neural network models aimed at abstracting, emulating, and adequately explaining the mechanisms of central visual object recognition in primates. Based on a wealth of primate neurophysiological and behavioral data, some of these network models are currently the most advanced (i.e., the most accurate) scientific theories of the internal mechanisms of primate ventral visual flow and how these mechanisms underpin the ability of humans and other primates to rapidly and accurately infer latent world content (e.g., object identity, position, pose, etc.) from the pixel set of most natural images. Although still far from complete, these cutting-edge scientific models already have many uses in brain science and beyond, and I will describe three recent examples from our team. First, I will describe recent neural and behavioral work comparing and contrasting object perception by primates with object perception by models in the context of an adversarial attack. Second, I will describe recent behavioral work comparing and contrasting object perception by primates with object perception by models in the context of rapid object learning. Third, if time permits, I will highlight the use of leading models to design patterns of light energy in the retina (i.e., personalized synthetic images) to precisely modulate neural activity deep in the brain. In our view, this is an exciting new avenue of potential clinical benefit to humans.
Speaker: Jim DiCarlo is Professor of Systems and Computational Neuroscience at the Massachusetts Institute of Technology. The main goal of his research team is to discover and artificially emulate the brain mechanisms underlying human visual intelligence. Over the past 20 years, DiCarlo and his collaborators have helped develop, using the non-human primate animal model organism, our contemporary engineering-level understanding of the neural mechanisms underlying the processing of visual information in the ventral visual stream – a complex series of interconnected brain areas – and how that processing underpins basic cognitive abilities such as object and face recognition. He and his collaborators aim to use these new scientific insights to guide the development of more robust computer vision (“AI”) systems, reveal new ways to beneficially modulate brain activity through modulating the images that reach our eyes, expose new methods to accelerate visual learning, lay the groundwork for new neural prostheses (brain-machine interfaces) to restore lost senses, and provide a scientific basis for understanding how sensory processing is altered in conditions such as agnosia, autism, and dyslexia. DiCarlo trained in biomedical engineering, medicine, systems neurophysiology and computer science at Northwestern (BSE), Johns Hopkins (MD/PhD) and Baylor College of Medicine (Postdoc). He was Director of the MIT Department of Cognitive and Brain Sciences from 2012-2021, and is currently Director of the MIT Quest for Intelligence (2021-present), where he and his leadership team work to advance interdisciplinary research at the interface of natural and artificial intelligence. DiCarlo is an Alfred P. Sloan Research Fellow, a Pew Scholar in Biomedical Sciences, a McKnight Scholar in Neuroscience, and an elected member of the American Academy of Arts & Sciences.
Recording: Reverse engineering of human brain mechanisms of object perception and learning
Quantum Machine Learning in High Energy Physics
Date: 16/01/2024
Abstract: Theoretical and algorithmic advances, availability of data, and computing power have opened the door to exceptional perspectives for application of classical Machine Learning in the most diverse fields of science, business and society at large, and notably in High Energy Physics (HEP). In particular, Machine Learning is among the most promising approaches to analyse and understand the data the next generation HEP detectors will produce. Machine Learning is also a promising task for near-term quantum devices that can leverage compressed high dimensional representations and use the stochastic nature of quantum measurements as random source. Several architectures are being investigated. Quantum implementations of Boltzmann Machines, classifiers or Auto-Encoders, among the most popular classical approaches, are being proposed for different applications. Born machines are purely quantum models that can generate probability distributions in a unique way, inaccessible to classical computers. This talk will give an overview of the current state of the art in terms of Machine Learning on quantum computers with focus on their application to HEP.
Speaker: Sofia Vallecorsa is a CERN physicist with extensive experience in software development in the high-energy physics domain, particularly in deep learning and quantum computing applications within CERN openlab (https://openlab.cern/index.php/). She has a PhD in physics obtained at the University of Geneva. Prior to joining CERN openlab, Sofia was responsible for the development of deep-learning-based technologies for the simulation of particle transport through detectors at CERN. She also worked to optimise the GeantV detector simulation prototype on modern hardware architectures. See https://scholar.google.com/citations?user=OQpf9YsAAAAJ&hl=en for an extensive list of publications.
DaSCI Seminars 2023
Deep Learning 2.0: Towards AI that Builds and Improves AI
Date: 10/10/2023
Abstract: Throughout the history of AI, there is a clear pattern that manual elements of AI methods are eventually replaced by better-performing automatically-found ones; for example, deep learning (DL) replaced manual feature engineering with learned representations. The logical next step in representation learning is to also (meta-)learn the best architectures for these representations, as well as the best algorithms & hyperparameters for learning them. In this talk, I will discuss various works with this goal in the area of AutoML, highlighting that AutoML can be efficient and arguing for an emphasis on multi-objective AutoML to also account for the various dimensions of trustworthiness (such as algorithmic fairness, robustness, and uncertainty calibration). Finally, taking the idea of meta-learning to the extreme, I will deep-dive into a novel approach that learns an entire classification algorithm for small tabular datasets that achieves a new state of the art at the cost of a single forward pass.
Speaker: Frank Hutter is a Full Professor for Machine Learning at the University of Freiburg (Germany). He holds a PhD from the University of British Columbia (UBC, 2009), for which he received the 2010 CAIAC doctoral dissertation award for the best thesis in AI in Canada. He also won several best paper awards and prizes in international ML competitions. He is a Fellow of ELLIS and EurAI, Director of the ELLIS unit Freiburg, and the recipient of 3 ERC grants. Frank is best known for his research on automated machine learning (AutoML), including neural architecture search, efficient hyperparameter optimization, and meta-learning. He co-authored the first book on AutoML and the prominent AutoML tools Auto-WEKA, Auto-sklearn and Auto-PyTorch, won the first two AutoML challenges with his team, is co-teaching the first MOOC on AutoML, co-organized 15 AutoML-related workshops at ICML, NeurIPS and ICLR, and founded the AutoML conference as general chair in 2022.
Data-driven models for efficient healthcare delivery
Date: 26/09/2023
Abstract: This talk will cover several projects where we combined Machine Learning and Optimization to treat more patients, using productivity-driven scheduling to reduce delays in dynamic /online contexts and to reduce treatment time for cancer care. We will cover projects collaborating with emergency departments, radiology, radio-oncology, and ambulance services.
Speaker: ouis-Martin Rousseau is a Professor in the Department of Mathematics and Industrial Engineering at École Polytechnique de Montréal. Since 2016, he has held the Canada Research Chair in Healthcare Analytics and Logistics, which studies complex and interconnected problems in-home care services, cancer treatment, and hospital logistics. Louis-Martin was also the founder and Chief Science Officer of Planora before its acquisition by JDA in 2012, where he served as a scientific advisor afterward. With students and colleagues, he recently cofounded Gray Oncology Solution (www.gray-os.com), which proposes a patient-scheduling SaaS solution to the health sector.
Recording: Data-driven models for efficient healthcare delivery
Machine learning in medicine: Sepsis prediction and antibiotic resistance prediction
Date: 13/06/2023
Abstract: Sepsis is a major cause of mortality in intensive care units around the world. If recognized early, it can often be treated successfully, but early prediction of sepsis is an extremely difficult task in clinical practice. The data wealth from intensive care units that is increasingly becoming available for research now allows to study this problem of predicting sepsis using machine learning and data mining approaches. In this talk, I will describe our efforts towards data-driven early recognition of sepsis and the related problem of antibiotic resistance prediction.
Speaker: Karsten Borgwardt is Director of the Department of Machine Learning and Systems Biology at the Max Planck Institute of Biochemistry in Martinsried, Germany since February 2023. His work won several awards, including the 1 million Euro Krupp Award for Young Professors in 2013 and a Starting Grant 2014 from the ERC-backup scheme of the Swiss National Science Foundation. Prof. Borgwardt has been leading large national and international research consortia, including the “Personalized Swiss Sepsis Study” (2018-2023) and the subsequent National Data Stream on infection-related outcomes in Swiss ICUs (2022-2023), and two Marie Curie Innovative Training Networks on Machine Learning in Medicine (2013-2016 and 2019-2022).
Recording: Machine learning in medicine: Sepsis prediction and antibiotic resistance prediction
Multiscale Random Models of Deep Neural Networks
Date: 16/05/2023
Abstract: Deep neural networks have spectacular applications but remain mostly a mathematical mystery. An outstanding issue is to understand how they circumvent the curse of dimensionality to generate or classify data. Inspired by the renormalization group in physics, we explain how deep networks can separate phenomena which appear at different scales, and capture scale interactions. It provides high-dimensional model, which approximate the probability distribution of complex physical fields such as turbulences or structured images. For classification, learning becomes similar to a compressed sensing problem, where low-dimensional discriminative structures are identified with random projections. We introduce a multiscale random feature model of deep networks for classification, which is validated numerically.
Speaker: Stéphane Mallat was Professor at NYU in computer science, until 1994, then at Ecole Polytechnique in Paris and Department Chair. From 2001 to 2007 he was co-founder and CEO of a semiconductor start-up company. Since 2017, he holds the “Data Sciences” chair at the Collège de France. He is a member of the French Academy of sciences, of the Academy of Technologies, and a foreign member of the US National Academy of Engineering. Stéphane Mallat’s research interests include machine learning, signal processing and harmonic analysis. He developed the multiresolution wavelet theory and algorithms at the origin of the compression standard JPEG-2000, and sparse signal representations in dictionaries through matching pursuits. He currently works on mathematical models of deep neural networks, for data analysis and physics.
Geometric Deep Learning: Grids, Graphs, Groups, Geodesics and Gauges
Date: 28/03/2023
Abstract:
The last decade has witnessed an experimental revolution in data science and machine learning, epitomised by deep learning methods. Indeed, many high-dimensional learning tasks previously thought to be beyond reach –such as computer vision, playing Go, or protein folding – are in fact feasible with appropriate computational scale. Remarkably, the essence of deep learning is built from two simple algorithmic principles: first, the notion of representation or feature learning, whereby adapted, often hierarchical, features capture the appropriate notion of regularity for each task, and second, learning by local gradient-descent type methods, typically implemented as backpropagation.
While learning generic functions in high dimensions is a cursed estimation problem, most tasks of interest are not generic, and come with essential pre-defined regularities arising from the underlying low-dimensionality and structure of the physical world. This talk is concerned with exposing these regularities through unified geometric principles that can be applied throughout a wide spectrum of applications.
Such a ‘geometric unification’ endeavour in the spirit of Felix Klein’s Erlangen Program serves a dual purpose: on one hand, it provides a common mathematical framework to study the most successful neural network architectures, such as CNNs, RNNs, GNNs, and Transformers. On the other hand, it gives a constructive procedure to incorporate prior physical knowledge into neural architectures and provide principled way to build future architectures yet to be invented.
Speaker: Petar Veličković (see https://petar-v.com/ for a short bio)
Recording: Geometric Deep Learning: Grids, Graphs, Groups, Geodesics and Gauges
Model-free, Model-based, and General Intelligence: Learning Representations for Acting and Planning
Date: 07/03/2023
Abstract: During the 60s and 70s, AI researchers explored intuitions about intelligence by writing programs that displayed intelligent behavior. Many good ideas came out from this work but programs written by hand were not robust or general. After the 80s, research increasingly shifted to the development of learners capable of inferring behavior and functions from experience and data, and solvers capable of tackling well-defined but intractable models like SAT, classical planning, Bayesian networks, and POMDPs. The learning approach has achieved considerable success but results in black boxes that do not have the flexibility, transparency, and generality of their model-based counterparts. Model-based approaches, on the other hand, require models and scalable algorithms. The two have close parallels with Daniel Kahneman’s Systems 1 and 2: the first, a fast, opaque, and inflexible intuitive mind; the second, a slow, transparent, and flexible analytical mind. In this talk, I review learners and solvers, and the challenge of integrating their System 1 and System 2 capabilities, focusing then on our recent work aimed at bridging this gap in the context of action and planning, where combinatorial and deep learning approaches are used to learn general action models, general policies, and general subgoal structures.
Speaker: Hector Geffner is an Alexander Humbolt Professor at RWTH Aachen University, Germany, and a Wallenberg Guest Professor at Linköping University, Sweden. Hector grew up in Buenos Aires and obtained a PhD in Computer Science at UCLA in 1989. He then worked at the IBM T.J. Watson Research Center in New York, at the Universidad Simon Bolivar in Caracas, and at the Catalan Institute of Advanced Research (ICREA) and the Universitat Pompeu Fabra in Barcelona. Hector teaches courses on logic, AI, and social and technological change, and is currently doing research on representations learning for acting and planning as part of the ERC project RLeap 2020-2025.
Recording: Model-free, Model-based, and General Intelligence: Learning Representations for Acting and Planning
DaSCI Seminars 2022
With a little help from NLP: My Language Technology applications with impact on society
Date: 21/11/2022
Abstract: The talk will present original methodologies developed by the speaker, underpinning implemented Language Technology tools which are already having an impact on the following areas of society: e-learning, translation and interpreting and care for people with language disabilities. The first part of the presentation will introduce an original methodology and tool for generating multiple-choice tests from electronic textbooks. The application draws on a variety of Natural Language Processing (NLP) techniques which include term extraction, semantic computing and sentence transformation. The presentation will include an evaluation of the tool which demonstrates that generation of multiple-choice tests items with the help of this tool is almost four times faster than manual construction and the quality of the test items is not compromised. This application benefits e-learning users (both teachers and students) and is an example of how NLP can have a positive societal impact, in which the speaker passionately believes. The latest version of the system based on deep learning techniques will also be briefly introduced. The talk will go on to discuss two other original recent projects which are also related to the application of NLP beyond academia. First, a project, whose objective is to develop next-generation translation memory tools for translators and, in the near future, for interpreters, will be briefly presented. Finally, an original methodology and system will be outlined which helps users with autism to read and better understand texts. The presentation will finish with a brief outline of the latest (and forthcoming) research topics (to be) pursued by the speaker and his vision on the future NLP applications.
Speaker: Prof Dr Ruslan Mitkov has been working in Natural Language Processing (NLP), Computational Linguistics, Corpus Linguistics, Machine Translation, Translation Technology and related areas since the early 1980s. Whereas Prof Mitkov is best known for his seminal contributions to the areas of anaphora resolution and automatic generation of multiple-choice tests, his extensively cited research (more than 270 publications including 20 books, 35 journal articles and 40 book chapters) also covers topics such as deep learning for NLP, machine translation, translation memory and translation technology in general, bilingual term extraction, automatic identification of cognates and false friends, natural language generation, automatic summarisation, computer-aided language processing, centering, evaluation, corpus annotation, NLP-driven corpus-based study of translation universals, text simplification, NLP for people with language disorders and computational phraseology. In addition, Ruslan Mitkov is well known for his vision in research based on innovative ideas and drive towards research output which seeks to enhance the work efficiency of different professions (e.g. for teachers, translators and interpreters) or seeks to improve the quality of life (e.g. for people with language disabilities) and which has significant impact beyond academia. Mitkov is author of the monograph Anaphora resolution (Longman) and Editor of the most successful Oxford University Press Handbook – The Oxford Handbook of Computational Linguistics whose second and substantially revised edition was published in June 2022.
Recording: With a little help from NLP: My Language Technology applications with impact on society
Trustworthy Learning in Heterogeneous Networks
Date: 08/11/2022
Abstract: Federated learning introduces a number of challenges beyond traditional distributed learning scenarios. In addition to being accurate, federated methods must scale to potentially massive and heterogeneous networks of devices, and must exhibit trustworthy behavior—addressing pragmatic concerns related to issues such as fairness, robustness, and user privacy. In this talk, I talk about how heterogeneity lies at the center of the constraints of federated learning—not only affecting the accuracy of the models, but also competing with other critical metrics such as fairness, robustness, and privacy. To address these metrics, I talk about new, scalable federated learning objectives and algorithms that rigorously account for and address sources of heterogeneity. Although our work is grounded by the application of federated learning, I show that many of the techniques and fundamental tradeoffs extend well beyond this use-case.
Speaker: Tian Li is a fifth-year Ph.D. student in the Computer Science Department at Carnegie Mellon University working with Virginia Smith. Her research interests are in distributed optimization, large-scale machine learning, federated learning, and data-intensive systems. Prior to CMU, she received her undergraduate degrees in Computer Science and Economics from Peking University. She was a research intern at Google Research in 2022. She received the Best Paper Award at ICLR Workshop on Security and Safety in Machine Learning Systems (2021), was selected as Rising Stars in Machine Learning by UMD (2021), Rising Stars in Data Science by UChicago (2022), and was invited to participate in EECS Rising Stars Workshop (2022).
Tackling climate change with machine learning
Date: 18/10/2022
Abstract: Machine learning can be a powerful tool in helping society reduce greenhouse gas emissions and adapt to a changing climate. In this talk, we will explore opportunities and challenges in AI-for-climate, from optimizing electrical grids to monitoring crop yield, and how methodological innovations in machine learning can be driven by impactful climate-relevant problems.
Speaker: David Rolnick is an Assistant Professor and Canada CIFAR AI Chair in the School of Computer Science at McGill University and at Mila Quebec AI Institute. He is a Co-founder and Chair of Climate Change AI and serves as Scientific Co-director of Sustainability in the Digital Age. Dr. Rolnick received his Ph.D. in Applied Mathematics from MIT. He is a former NSF Mathematical Sciences Postdoctoral Research Fellow, NSF Graduate Research Fellow, and Fulbright Scholar, and was named to the MIT Technology Review’s 2021 list of “35 Innovators Under 35.”
Coevolutionary Probelm-Solving
Date: 04/10/2022
Abstract: Coevolution is an old but very interesting research topic in evolutionary computation. This talk presents some of the applications of coevolution in learning and optimisation. First, we look at a classical coevolutionary learning scenario when no training data are available. In fact, no teacher information is available either. Then we examine how coevolution could be used to tackle large-scale global optimisation in the black-box optimisation setting. Finally, we explore how coevolution could be harnessed to design general solvers automatically for hard combinatorial optimisation problems.
Speaker: Xin Yao is a Chair Professor of Computer Science at the Southern University of Science and Technology (SUSTech), Shenzhen, China, and a part-time Professor of Computer Science at the University of Birmingham, UK. He is an IEEE Fellow and was a Distinguished Lecturer of the IEEE Computational Intelligence Society (CIS). He served as the President (2014-15) of IEEE CIS and the Editor-in-Chief (2003-08) of IEEE Transactions on Evolutionary Computation. His major research interests include evolutionary computation, ensemble learning, and their applications to software engineering. His research work won the 2001 IEEE Donald G. Fink Prize Paper Award; 2010, 2016 and 2017 IEEE Transactions on Evolutionary Computation Outstanding Paper Awards; 2011 IEEE Transactions on Neural Networks Outstanding Paper Award; and many other best paper awards at conferences. He received a Royal Society Wolfson Research Merit Award in 2012, the IEEE CIS Evolutionary Computation Pioneer Award in 2013 and the 2020 IEEE Frank Rosenblatt Award.
Recording: Coevolutionary Probelm-Solving
Evolutionary Intelligent for Data Analytics and Optimization
Date: 13/07/2022
Abstract: Evolutionary Intelligence (EI) has been widely used during the last two decades and has remained a highly-researched topic, especially for complex real-world problems. The EI techniques are a subset of artificial intelligence, but they are slightly different from the classical methods in the sense that the intelligence of EI comes from biological systems or nature in general. The efficiency of EC is due to their significant ability to imitate the best features of nature which have evolved by natural selection over millions of years. The central theme of this presentation is about EI techniques and their application to complex real-world engineering problems. On this basis, first I will talk about an automated learning approach called genetic programming. Applied evolutionary learning will be presented, and then their new advances will be mentioned. Here, some of my studies on big data analytics and modelling using EI and genetic programming, in particular, will be presented. Second, evolutionary optimization will be presented including key applications in the design optimization of complex and nonlinear engineering systems. It will also be explained how such algorithms have been adopted to engineering problems and how their advantages over the classical optimization problems are used in action. Optimization results of large-scale towers and many-objective problems will be presented which show the applicability of EI. Finally, heuristics will be explained which are adaptable with EI and they can significantly improve the optimization results.
Speaker: Amir H. Gandomi is a Professor of Data Science and an ARC DECRA Fellow at the Faculty of Engineering & Information Technology, University of Technology Sydney. Prior to joining UTS, Prof. Gandomi was an Assistant Professor at Stevens Institute of Technology, USA and a distinguished research fellow at BEACON center, Michigan State University, USA. Prof. Gandomi has published over three hundred journal papers and 12 books which collectively have been cited 30,000+ times (H-index = 79). He has been named as one of the most influential scientific minds and Highly Cited Researcher (top 1% publications and 0.1% researchers) for five consecutive years, 2017 to 2021. He also ranked 17th in GP bibliography among more than 12,000 researchers. He has received multiple prestigious awards for his research excellence and impact, such as the 2022 Walter L. Huber Prize which is known as the highest level mid-career research award in all areas of civil engineering. He has served as associate editor, editor, and guest editor in several prestigious journals such as AE of IEEE TBD and IEEE IoTJ. Prof Gandomi is active in delivering keynotes and invited talks. His research interests are global optimisation and (big) data analytics using machine learning and evolutionary computations in particular.
Recording: Evolutionary Intelligent for Data Analytics and Optimization
Autonomous Driving Research at CVC/UAB
Date: 06/07/2022
Abstract: Developing autonomous vehicles is a complex challenge. It involves training and testing AI drivers by using supervised data collected on a diversity of driving episodes. We could say that data is the driver in autonomous driving, especially in the deep learning era. In this context, the talk focuses on the efforts conducted at CVC to minimize data labelling efforts. This includes the use of simulated data to support the training of visual models, the development of self-labeling procedures, as well as exploring non-standard paradigms for autonomous driving such as end-to-end driving by imitation learning.
Speaker: Antonio M López has a long trajectory carrying research at the intersection of computer vision, simulation, machine learning, driver assistance, and autonomous driving. Antonio has been deeply involved in the creation of the SYNTHIA dataset and the CARLA open-source simulator, both for democratizing autonomous driving research. He is actively working hand-on-hand with industry partners to bring state-of-the-art techniques to the field of autonomous driving.
Recording: Autonomous Driving Research at CVC/UAB
Neurosymbolic Computing for Accountability in AI
Date: 11/05/2022
Abstract: Despite achieving much success, the deep learning approach to AI has been criticised for being “black box”: the decisions made by such large and complex learning systems are difficult to explain or analyse. If the system makes a mistake in a critical situation then the consequences can be serious. The use of black box systems has obvious implications to transparency but also fairness and ultimately trust in current AI. System developers might also like to learn from system errors so that errors can be fixed. The area of explainable AI (XAI) has sought to open the black box by providing explanations for large AI systems mostly through the use of visualization techniques and user studies that seek to associate the decisions made by the system with known features of the deep learning model. In this talk, I will argue that XAI needs knowledge extraction and an objective measure of fidelity as a pre-requisite for visualization and user studies. As part of a neurosymbolic approach, knowledge extraction creates a bridge between sub-symbolic deep learning and logic-based symbolic AI with a precise semantics. I will exemplify how knowledge extraction can be used in the analysis of chest x-ray images as part of a collaborative project with Fujitsu Research to find and fix mistakes in image classification. I will conclude by arguing that knowledge extraction is an important tool, but is only one of many elements that are needed to address fairness and accountability in AI.
Speaker: Artur Garcez is Professor of Computer Science and Director of the Data Science Institute at City, University of London. He holds a PhD in Computing (2000) from Imperial College London. He is a Fellow of the British Computer Society (FBCS) and president of the steering committee of the Neural-Symbolic Learning and Reasoning Association. He has co-authored two books: Neural-Symbolic Cognitive Reasoning, 2009, and Neural-Symbolic Learning Systems, 2002. His research has led to publications in the journals Behavioral & Brain Sciences, Theoretical Computer Science, Neural Computation, Machine Learning, Journal of Logic and Computation, IEEE Transactions on Neural Networks, Journal of Applied Logic, Artificial Intelligence, and Studia Logica, and the flagship AI and Neural Computation conferences AAAI, NeurIPS, IJCAI, IJCNN, AAMAS and ECAI. Professor Garcez holds editorial positions with several scientific journals in the fields of Computational Logic and Artificial Intelligence, and has been Programme Committee member for several conferences, including IJCAI, IJCNN, NeurIPS and AAAI.
The Modern Mathematics of Deep Learning
Date: 05/04/2022
Abstract: Despite the outstanding success of deep neural networks in real-world applications, ranging from science to public life, most of the related research is empirically driven and a comprehensive mathematical foundation is still missing. At the same time, these methods have already shown their impressive potential in mathematical research areas such as imaging sciences, inverse problems, or numerical analysis of partial differential equations, sometimes by far outperforming classical mathematical approaches for particular problem classes. The goal of this lecture is to first provide an introduction into this new vibrant research area. We will then survey recent advances in two directions, namely the development of a mathematical foundation of deep learning and the introduction of novel deep learning-based approaches to solve mathematical problem settings.
Speaker: Gitta Kutyniok (https://www.ai.math.lmu.de/kutyniok) currently has a Bavarian AI Chair for Mathematical Foundations of Artificial Intelligence at the Ludwig-Maximilians Universität München. She received her Diploma in Mathematics and Computer Science as well as her Ph.D. degree from the Universität Paderborn in Germany, and her Habilitation in Mathematics in 2006 at the Justus-Liebig Universität Gießen. From 2001 to 2008 she held visiting positions at several US institutions, including Princeton University, Stanford University, Yale University, Georgia Institute of Technology, and Washington University in St. Louis, and was a Nachdiplomslecturer at ETH Zurich in 2014. In 2008, she became a full professor of mathematics at the Universität Osnabrück, and moved to Berlin three years later, where she held an Einstein Chair in the Institute of Mathematics at the Technische Universität Berlin and a courtesy appointment in the Department of Computer Science and Engineering until 2020. In addition, Gitta Kutyniok holds an Adjunct Professorship in Machine Learning at the University of Tromso since 2019.
Neuroevolution: A Synergy of Evolution and Learning
Date: 22/03/2022
Abstract: Neural network weights and topologies were originally evolved in order to solve tasks where gradients are not available. Recently, it has also become a useful technique for metalearning architectures of deep learning networks. However, neuroevolution is most powerful when it utilizes synergies of evolution and learning. In this talk I review several examples of such synergies: evolving loss functions, activation functions, surrogate optimization, and human-designed solutions. I will demonstrate these synergies in image recognition, game playing, and pandemic policy optimization, and point out opportunities for future work.
Speaker: Risto Miikkulainen is a Professor of Computer Science at the University of Texas at Austin and Associate VP of Evolutionary AI at Cognizant. He received an M.S. in Engineering from Helsinki University of Technology (now Aalto University) in 1986, and a Ph.D. in Computer Science from UCLA in 1990. His current research focuses on methods and applications of neuroevolution, as well as neural network models of natural language processing and vision; he is an author of over 450 articles in these research areas. At Cognizant, he is scaling up these approaches to real-world problems. Risto is an IEEE Fellow; his work on neuroevolution has recently been recognized with the IEEE CIS Evolutionary Computation Pioneer Award, the Gabor Award of the International Neural Network Society and Outstanding Paper of the Decade Award of the International Society for Artificial Life.
Trustable autonomy: creating interfaces between human and robot societies
Date: 26/01/2022
Abstract: Robotic systems are starting to revolutionize many applications, from transportation to health care, assisted by technological advancements, such as cloud computing, novel hardware design, and novel manufacturing techniques. However, several of the characteristics that make robots ideal for certain future applications such as autonomy, self-learning, knowledge sharing, can also raise concerns in the evolution of the technology from academic institutions to the public sphere. Blockchain, an emerging technology originated in the digital currency field, is starting to show great potential to make robotic operations more secure, autonomous, flexible, and even profitable. Therefore, bridging the gap between purely scientific domains and real-world applications. This talk seeks to move beyond the classical view of robotic systems to advance our understanding about the possibilities and limitations of combining state-of-the art robotic systems with blockchain technology.
Speaker: Eduardo Castello experience and interests comprise robotics, blockchain technology, and complex systems. Eduardo was a Marie Curie Fellow at the MIT Media Lab where he worked to explore the combination of distributed robotic systems and blockchain technology. His work focuses on implementing new security, behavior, and business models for distributed robotics by using novel cryptographic methods. Eduardo received his Bsc.(Hons) intelligent systems from University of Portsmouth (UK) and his M. Eng and Ph.D degrees in robotics engineering from Osaka University (Japan). During his graduate studies, Eduardo’s research focused on swarm robotics and how to achieve cooperative and self-sustaining groups of robots.
DaSCI Seminars 2021
If all you have is a hammer, everything looks like a nail
Date: 01/12/2021
Abstract: In this talk, I’ll focus on some recent advances in privacy-preserving NLP. In particular, we will look at the differential privacy paradigm and its applications in NLP, namely by using differentially-private training of neural networks. Although the training framework is very general, does it really fit everything we typically do in NLP?
Speaker: Dr. Habernal is leading an independent research group “Trustworthy Human Language Technologies” at the Department of Computer Science, Technical University of Darmstadt, Germany. His current research areas include privacy-preserving NLP, legal argument mining, and explainable and trustworthy models. His research track spans argument mining and computational argumentation, crowdsourcing, or serious games, among others. More info at www.trusthlt.org.
Graph Mining with Graph Neural Networks
Lecturer: Bryan Perozzi is a Research Scientist in Google Research’s Algorithms and Optimization group, where he routinely analyzes some of the world’s largest (and perhaps most interesting) graphs. Bryan’s research focuses on developing techniques for learning expressive representations of relational data with neural networks. These scalable algorithms are useful for prediction tasks (classification/regression), pattern discovery, and anomaly detection in large networked data sets. Bryan is an author of 30+ peer-reviewed papers at leading conferences in machine learning and data mining (such as NeurIPS, ICML, KDD, and WWW). His doctoral work on learning network representations was awarded the prestigious SIGKDD Dissertation Award. Bryan received his Ph.D. in Computer Science from Stony Brook University in 2016, and his M.S. from the Johns Hopkins University in 2011.
Date:17/05/2021
Abstract: How can neural networks best model data which doesn’t have a fixed structure? In this talk, I will discuss graph neural networks (GNNs), a very active area of current research in machine learning aimed at answering this interesting (and practical) question. After reviewing the basics of GNNs, I’ll discuss some challenges applying these methods in industry, and some of the methods we’ve developed for addressing these challenges.
Detecting the “Fake News” Before It Was Even Written, Media Literacy, and Flattening the Curve of the COVID-19 Infodemic
Lecturer: Dr. Preslav Nakov is a Principal Scientist at the Qatar Computing Research Institute (QCRI), HBKU , where he leads the Tanbih mega-project (developed in collaboration with MIT ), which aims to limit the effect of “fake news”, propaganda and media bias by making users aware of what they are reading, thus promoting media literacy and critical thinking. He received his PhD degree in Computer Science from the University of California at Berkeley, supported by a Fulbright grant. Dr. Preslav Nakov is President of ACL SIGLEX , Secretary of ACL SIGSLAV , and a member of the EACL advisory board. He is also member of the editorial board of a number of journals including Computational Linguistics, TACL , CS&L, NLE , AI Communications, and Frontiers in AI. He authored a Morgan & Claypool book on Semantic Relations between Nominals and two books on computer algorithms. He published 250+ research papers, and he was named among the top 2% of the world’s most-cited in the career achievement category, part of a global list compiled by Stanford University. He received a Best Long Paper Award at CIKM ‘2020, a Best Demo Paper Award (Honorable Mention) at ACL ‘2020, a Best Task Paper Award (Honorable Mention) at SemEval’2020, a Best Poster Award at SocInfo’2019, and the Young Researcher Award at RANLP ‘2011. He was also the first to receive the Bulgarian President’s John Atanasoff award, named after the inventor of the first automatic electronic digital computer. Dr. Nakov’s research was featured by over 100 news outlets, including Forbes, Boston Globe, Aljazeera, DefenseOne, Business Insider, MIT Technology Review, Science Daily, Popular Science, Fast Company, The Register, WIRED , and Engadget, among others.
Date:19/04/2021
Abstract: Given the recent proliferation of disinformation online, there has been growing research interest in automatically debunking rumors, false claims, and “fake news”. A number of fact-checking initiatives have been launched so far, both manual and automatic, but the whole enterprise remains in a state of crisis: by the time a claim is finally fact-checked, it could have reached millions of users, and the harm caused could hardly be undone.
An arguably more promising direction is to focus on analyzing entire news outlets, which can be done in advance; then, we could fact-check the news before it was even written: by checking how trustworthy the outlet that has published it is (which is what journalists actually do). We will show how we do this in the Tanbih news aggregator (http://www.tanbih.org/), which aims to limit the impact of “fake news”, propaganda and media bias by making users aware of what they are reading, thus promoting media literacy and critical thinking, which are arguably the best way to address disinformation in the long run. In particular, we develop media profiles that show the general factuality of reporting, the degree of propagandistic content, hyper-partisanship, leading political ideology, general frame of reporting, stance with respect to various claims and topics, as well as audience reach and audience bias in social media.
Another important observation is that the term “fake news” misleads people to focus exclusively on factuality, and to ignore the other half of the problem: the potential malicious intent. Thus, we detect the use of specific propaganda techniques in text, e.g., appeal to emotions, fear, prejudices, logical fallacies, etc. We will show how we do this in the Prta system (https://www.tanbih.org/prta), another media literacy tool, which got the Best Demo Award (Honorable Mention) at ACL -2020; an associated shared task got the Best task award (Honorable Mention) at SemEval-2020.
Finally, at the time of COVID -19, the problem of disinformation online got elevated to a whole new level as the first global infodemic. While fighting this infodemic is typically thought of in terms of factuality, the problem is much broader as malicious content includes not only “fake news”, rumors, and conspiracy theories, but also promotion of fake cures, panic, racism, xenophobia, and mistrust in the authorities, among others. Thus, we argue for the need of a holistic approach combining the perspectives of journalists, fact-checkers, policymakers, social media platforms, and society as a whole, and we present our recent research in that direction (https://mt.qcri.org/covid19disinformationdetector/).
Efficient Deep Learning
Lecturer: Marco Pedersoli is Assistant Professor at ETS Montreal. He obtained his PhD in computer science in 2012 at the Autonomous University of Barcelona and the Computer Vision Center of Barcelona. Then, he was a postdoctoral fellow in computer vision and machine learning at KU Leuven with Prof. Tuytelaars and later at INRIA Grenoble with Drs. Verbeek and Schmid. At ETS Montreal he is a member of LIVIA and he is co-chairing an industrial Chair on Embedded Neural Networks for Connected Building Control. His research is mostly applied on visual recognition, the automatic interpretation and understanding of images and videos. His specific focus is on reducing the complexity and the amount of annotation required for deep learning algorithms such as convolutional and recurrent neural networks. Prof. Pedersoli has authored more than 40 publications in top-tier international conferences and journals in computer vision and machine learning.
Date:12/04/2021
Abstract: In the last 10 years deep learning (DL) models have shown great progress in many different fields, from Computer Vision to Natural Language Processing. However, DL methods require great computational resources (i.e. GPUs or TPUs) and very large datasets, which also makes the training phase very long and painful. Thus, there is a strong need for reducing the computational cost of DL methods both in training as well as in deployment. In this talk, I present the most common families of approaches used to reduce the requirements of DL methods in terms of Memory and Computation for both training and deployment, and show how a reduction of the model footprint does not always produce a corresponding speed-up. Finally, I will present some recent results that suggest that large DL models are important mostly for facilitating the model training, and when that is finished, we can deploy a much smaller and faster model with almost no loss in accuracy.
Variational Autoencoders for Audio, Visual and Audio-Visual Learning
Lecturer: Xavier Alameda-Pineda is a (tenured) Research Scientist at Inria, in the Perception Group. He obtained the M.Sc. (equivalent) in Mathematics in 2008, in Telecommunications in 2009 from BarcelonaTech and in Computer Science in 2010 from Université Grenoble-Alpes (UGA). He then worked towards his Ph.D. in Mathematics and Computer Science, and obtained it in 2013, from UGA. After a two-year post-doc period at the Multimodal Human Understanding Group, at University of Trento, he was appointed with his current position. Xavier is an active member of SIGMM, and a senior member of IEEE and a member of ELLIS. He is co-chairing the “Audio-visual machine perception and interaction for companion robots” chair of the Multidisciplinary Institute of Artificial Intelligence. Xavier is the Coordinator of the H2020 Project SPRING: Socially Pertinent Robots in Gerontological Healthcare. Xavier’s research interests are in combining machine learning, computer vision and audio processing for scene and behavior analysis and human-robot interaction. More info at xavirema.eu
Date: 01/02/2021
Abstract: Since their introduction, Variational Autoencoders (VAE) demonstrated great performance in key unsupervised feature representation applications, specifically in visual and auditory representation. In this seminar, the global methodology of variational auto-encoders will be presented, along with applications in learning with audio and visual data. Special emphasis will be put in discussing the use of VAE for audio-visual learning, showcasing its interest with the task of audio-visual speech enhancement.
Variational Autoencoders for Audio, Visual and Audio-Visual Learning – Recording
Five Sources of Biases and Ethical Issues in NLP, and What to Do about Them
Lecturer: Dirk Hovy is associate professor of computer science at Bocconi University in Milan, Italy. Before that, he was faculty and a postdoc in Copenhagen, got a PhD from USC, and a linguistics masters in Germany. He is interested in the interaction between language, society, and machine learning, or what language can tell us about society, and what computers can tell us about language. He has authored over 60 articles on these topics, including 3 best paper awards. He has organized one conference and several workshops (on abusive language, ethics in NLP, and computational social science). Outside of work, Dirk enjoys cooking, running, and leather-crafting. For updated information, see http://www.dirkhovy.com
Date: 11/01/2021
Abstract: Never before was it so easy to write a powerful NLP system, never before did it have such a potential impact. However, these systems are now increasingly used in applications they were not intended for, by people who treat them as interchangeable black boxes. The results can be simple performance drops, but also systematic biases against various user groups. In this talk, I will discuss several types of biases that affect NLP models (based on Shah et al. 2020 and Hovy & Spruit, 2016), what their sources are, and potential counter measures.
Five Sources of Biases and Ethical Issues in NLP, and What to Do about Them – Recording
Image and Video Generation using Deep Learning
Lecturer: Stéphane Lathuilière is an associate professor (maître de conférence) at Telecom Paris, France, in the multimedia team. Until October 2019, he was a post-doctoral fellow at the University of Trento (Italy) in the Multimedia and Human Understanding Group, led by Prof. Nicu Sebe and Prof. Elisa Ricci. He received the M.Sc. degree in applied mathematics and computer science from ENSIMAG, Grenoble Institute of Technology (Grenoble INP), France, in 2014. He completed his master thesis at the International Research Institute MICA (Hanoi, Vietnam). He worked towards his Ph.D. in mathematics and computer science in the Perception Team at Inria under the supervision of Dr. Radu Horaud, and obtained it from Université Grenoble Alpes (France) in 2018. His research interests cover machine learning for computer vision problems (eg. domain adaptation, continual learning) and deep models for image and video generation. He regularly publishes papers in the most prestigious computer vision conferences (CVPR, ICCV, ECCV, NeurIPS) and top journals (IEEE TPAMI).
Date: 14/12/2020
Abstract: Generating realistic images and videos has countless applications in different areas, ranging from photography technologies to e-commerce business. Recently, deep generative approaches have emerged as effective techniques for generation tasks. In this talk, we will first present the problem of pose-guided person image generation. Specifically, given an image of a person and a target pose, a new image of that person in the target pose is synthesized. We will show that important body-pose changes affect generation quality and that specific feature map deformations lead to better images. Then, we will present our recent framework for video generation. More precisely, our approach generates videos where an object in a source image is animated according to the motion of a driving video. In this task, we employ a motion representation based on keypoints that are learned in a self-supervised fashion. Therefore, our approach can animate any arbitrary object without using annotation or prior information about the specific object to animate.
Aggregating Weak Annotations from Crowds
Lecturer: Edwin Simpson, is a lecturer (equivalent to assistant professor) at the University of Bristol, working on interactive natural language processing. His research focusses on learning from small and unreliable data, including user feedback, and adapts Bayesian approaches to topics such as argumentation, summarisation and sequence labelling. Previously, he was a post-doc at TU Darmstadt, Germany and completed his PhD at the University of Oxford on Bayesian methods for aggregating crowdsourced data.
Date: 09/11/2020
Abstract: Current machine learning methods are data hungry. Crowdsourcing is a common solution to acquiring annotated data at large scale for a modest price. However, the quality of the annotations is highly variable and annotators do not always agree on the correct label for each data point. This talk presents techniques for aggregating crowdsourced annotations using preference learning and classifier combination to estimate gold-standard rankings and labels, which can be used as training data for ML models. We apply approximate Bayesian approaches to handle noise, small amounts of data per annotator, and provide a basis for active learning. While these techniques are applicable to any kind of data, we demonstrate their effectiveness for natural language processing tasks.
How to do Data Science without writing code
Lecturer: Victoriano Izquierdo (1990) is a software engineer from Granada who is the co-founder and CEO of Graphext. Graphext is developing an advanced data analysis software built upon the last advances in data science and artificial intelligence for supporting small and big companies to address hard problems using data.
Date: 15/02/2021
Abstract: How to do Data Science without writing code
DaSCI Seminars 2020
Reinforcement Learning
Lecturer: Sergio Guadarrama is a senior software engineer at Google Brain. He focuses on reinforcement learning and neural networks. His research focuses on robust, scalable and efficient reinforcement learning. He is currently the leader of the TF-Agents project and a lead developer of TensorFlow (co-creator of TF-Slim). Before joining Google, he was a researcher at the University of California, Berkeley, where he worked with Professor Lotfi Zadeh and Professor Trevor Darrell. He received his B.A. and Ph.D. from the Universidad Politécnica de Madrid.
Date: 26/10/2020
Abstract: Learning by reinforcement (RL) is a type of machine learning where the objective is to learn to solve a task through interaction with the environment, maximizing the expected return. Unlike supervised learning, the solution requires making multiple decisions in a sequential way and the reinforcement occurs through rewards. The two main components are the environment, which represents the problem to be solved, and the agent, which represents the learning algorithm.