Engineering UPF Seminars
Seminars take place at UPF, Campus Poblenou, Roc Boronat, 138, Barcelona and will only be streamed/recorded if the speaker has granted permission.
Rooms 55.309 streaming / 55.410 streaming / Auditorium streaming
If you are organizing a seminar, please fill in this form.
Upcoming
|
20 November 2026 12:30 · Room 52.217 Andrea Protani and Marc Molina van den Bosch PhD students at UPF, carrying out their doctoral research full-time at CERN through three-year doctoral appointments. |
Beyond the Final Paper: Building Shared, Traceable Research with Kosmico Research increasingly happens across disconnected folders, tools, messages, and AI chats. The final result may be a paper, but the reasoning that produced it, including hypotheses, experiments, decisions, and failed attempts, is often difficult to reconstruct. Kosmico starts from a simple idea: the project, not only the paper, should be the shared object of research. It brings literature, notes, code, experiments, writing, discussions, and AI agents into a single workspace, keeping the context and evolution of the project connected. In this hands-on seminar, participants will create a Kosmico project around their own research, explore its literature review and project management tools, and see how researchers and AI agents can work together without losing traceability. The session is especially intended for early-stage PhD students, but it is open to anyone interested. |
|
27 November 2026 11:00 · Room 52.223 Alexia Salavrakos Institute of Cosmos Sciences, University of Barcelona (ICCUB) |
Quantum machine learning, generative learning, and the train-on-classical deploy-on-quantum framework In this talk, I will do an introduction to the field of quantum machine learning (QML) and in particular to quantum generative learning. I will then explain how and why the train-on-classical deploy-on-quantum (TCDQ) framework was developed for generative QML. The TCDQ approach exploits the properties of so-called intermediate-complexity circuits, whose training can be simulated classically efficiently, but that generally require quantum hardware for the inference phase. Classical simulation of the training phase allows for automatic differentiation and models can be successfully trained up to thousands of qubits. Finally, I will show how quantum linear optical circuits can be used as generative QML models and how the TCDQ framework can be extended to these optical systems. |
|
3 December 2026 12:30 · Room 55.309 Bruno Galerne Professor of Mathematics · Denis Poisson Institute (CNRS UMR 7013), University of Orléans · Institut Universitaire de France (IUF) |
Diffusion models for Gaussian distributions: Exact solutions and Wasserstein errors Diffusion or score-based models recently showed high performance in image generation. They rely on a forward and a backward stochastic differential equations (SDE). The sampling of a data distribution is achieved by numerically solving the backward SDE or its associated flow ODE. Studying the convergence of these models necessitates to control four different types of error: the initialization error, the truncation error, the discretization error and the score approximation. In this work, we theoretically study the behavior of diffusion models and their numerical implementation when the data distribution is Gaussian. Our first contribution is to derive the analytical solutions of the backward SDE and the probability flow ODE and to prove that these solutions and their discretizations are all Gaussian processes. Our second contribution is to compute the exact Wasserstein errors between the target and the numerically sampled distributions for any numerical scheme. This allows us to monitor convergence directly in the data space, while experimental works limit their empirical analysis to Inception features. If time allows, extension of our approach for solving linear inverse problems will be discussed. |
Past seminars
Jump to year: 2026 · 2025 · 2024 · 2023
2026
|
22 September 2026 10:00 · Room 51.100 Anna Niarakis Full Professor of Bioinformatics and Quantitative Biology, University of Toulouse III-Paul Sabatier |
Large-scale mechanistic modeling for human complex diseases: From pathways to cells to systems The presentation will focus on developing mechanistic and hybrid digital twins to study complex immune-mediated diseases, with an emphasis on rheumatoid arthritis. I will discuss how multi-omics data can be integrated with prior biological knowledge to build executable, multiscale models that combine Boolean networks and agent-based modeling. The construction of digital twins for complex inflammatory and autoimmune diseases requires computational frameworks that integrate molecular regulation, cellular decision-making, and tissue-level dynamics. In this context, agent-based models (ABMs) offer a natural representation of spatially resolved immune cell interactions, whereas Boolean network models provide scalable and mechanistically grounded descriptions of intracellular signaling and cell fate control. |
|
15 September 2026 12:00 · Room 51.100 Alba Cervera |
Learning with and from Quantum Systems Information is physical. It is carried by some physical system, and what that system can and cannot do determines what can be done with the information it holds. It is natural, then, to ask what happens when the physical support is itself quantum, i.e. where the carrier can be placed in superposition, entangled, and described by complex amplitudes. This talk surveys my work and vision about what quantum information can offer in computation and how traditional computation can help us in designing new algorithms and applications. In particular, I will focus on the resulting two-way interface between quantum computing and machine learning. |
|
4 June 2026 16:00 · Room 55.309 Gemma Piella, Albert Hernansanz, Marc Balle UPF |
AI Across the Medical Continuum: From Perinatal Brain Imaging to Autonomous Surgery A three-part session. Gemma Piella — Computational Image Analysis for the Study of Early Brain Development: describing the basic components of an analysis pipeline that enables the quantification of perinatal brain development. Albert Hernansanz — on the shift from mechanical master-slave systems to intelligent, autonomous surgical agents capable of high-level cognitive decision-making and the role of eXplainable AI. Marc Balle — Vision-Language-Action (VLA) Models for Surgical Robotics: tracing the evolution of robotic autonomy and exploring the potential and challenges of bringing foundation models into the surgical operating room. |
|
6 May 2026 12:30 · Room 55.309 Ali Tajer Professor of Electrical, Computer, and Systems Engineering, RPI |
Causal Representation Learning Building reliable world models for physical systems is a central challenge in modern AI, with broad implications for scientific discovery, robotics, autonomy, and control. Unlike purely statistical prediction, world models for physical environments must capture the underlying causal structure of the system. We will discuss causal representation learning (CRL) as a principled foundation for building world models of physical systems, outline key challenges in learning causal latent structure from high-dimensional observations, and highlight emerging directions for integrating CRL with dynamics, control, and scientific machine learning. |
|
23 March 2026 15:00 · Room 55.309 Satoshi Oyama Professor, School of Data Science, Nagoya City University |
From Crowdsourcing to AI Agents: Collaborative Intelligence for Data Science and Scientific Discovery Modern data science relies not only on algorithms but also on large-scale collaboration among humans and machines. In this talk, I introduce research on crowdsourcing for data analytics, including statistical methods for integrating noisy crowd-generated data and competition platforms for predictive modeling. I then discuss how these ideas extend to emerging frameworks where AI agents participate in collaborative problem solving and scientific discovery. |
|
18 February 2026 16:30 · Room 55.309 Oscar Camara, Andy Luis Olivares, Manal Barrouhou UPF |
From Prehistoric PhDs to Computational Modelling in Cardiology A three-part session. Oscar Camara — Was a PhD different 20 years ago?: personal reflections comparing the PhD journey across two decades. Andy L. Olivares — Computational Modeling of Structural Deformation in LAAO and TAVI Cardiovascular Devices: a preliminary finite element model to evaluate deformation and device–tissue interaction. Manal Barrouhou — Modelling left atrial wall motion in atrial fibrillation and its impact on hemodynamics. |
2025
|
29 October 2025 10:00 · Room 55.309 Marcello Federico Senior Principal Scientist, Amazon Stores (EU INTech), Madrid |
Pioneering the Future of AI in E-commerce at Amazon Artificial Intelligence is reshaping industries globally, with e-commerce at the forefront of this transformation. This seminar delves into the cutting-edge work of EU INtech, Amazon Stores' European technology hub. We will explore the scientific directions guiding EU INtech's research and key application areas where AI is making significant impacts in e-commerce, the specific scientific challenges we're tackling, and Amazon's collaborative model with the broader scientific community. |
|
21 October 2025 16:00 · Room 55.309 Anirban Mukhopadhyay Independent Research Group Leader, TU Darmstadt |
Edible Foundation Models What if diagnosing cancer pan-Europe could be as simple as swallowing a capsule? Discover how cutting-edge Neural Cellular Automata (NCA) are transforming the future of healthcare — bringing high-accuracy diagnostics to smartphones, Raspberry Pis, and even ingestible capsules. Learn how this tech slashes energy use by 1000x, ensures privacy-preserving federated learning on smartphone CPUs, and breaks barriers in accessibility and equity. |
|
10 July 2025 10:00 · Room 55.410 Damien Vergnaud École normale supérieure |
Multi-Party Computation in the Head: Techniques and Applications Zero-knowledge proofs allow a prover to convince a verifier of the validity of a mathematical statement without revealing anything else. The paradigm of 'Multi-Party Computation in the Head' was introduced by Ishai, Kushilevitz, Ostrovsky, and Sahai in 2007 and has attracted significant interest in recent years. This approach notably enables the construction of efficient digital signature schemes whose security relies on the hardness of various algorithmic problems, including against potential quantum computers. |
|
3 July 2025 16:30 · Room 55.309 Inma Villanueva, Mané Sarkissian, Jorge Mateos UPF |
Mechanistic Analysis A three-part session. Inma Villanueva — Computational modeling of hemodynamic changes in fetuses with coarctation of the aorta. Mané Sarkissian — A Hybrid 3D Computational Model to Explore the Origins of Atherosclerosis. Jorge Mateos — Engineering Long Bone Growth: Computational Modeling of Growth Plate Mechanics. |
|
2 July 2025 16:00 · Room 55.309 Fangyijie Wang PhD Researcher, University College Dublin (ML-Labs) |
Low-Cost Deep Learning for Fetal Growth Monitoring in Low-Resource Settings Accurate fetal growth monitoring is essential for reducing maternal and neonatal complications, yet access to expert ultrasound interpretation remains limited in many low-resource settings. This talk presents recent efforts to develop low-cost deep learning methods for fetal ultrasound analysis, focusing on segmentation and gestational age estimation tasks, exploring lightweight architectures, semi-supervised learning techniques, and domain adaptation strategies. |
|
30 June 2025 12:00 · Room 52.321 Antonio Porras Associate Professor, University of Colorado Anschutz Medical Campus |
A data-driven approach to the clinical management of craniosynostosis The clinical management of patients with craniosynostosis is largely subjective. Consequently, patient outcomes are highly variable between institutions and suboptimal treatments are common. Dr. Porras will summarize his team's work creating image computing and data science methods specifically designed to study pediatric cranial development pathologies, and creating clinical translation protocols to objectify the evaluation and treatment of craniosynostosis. |
|
18 June 2025 16:00 · Room 55.309 Laura Becerra, Mar Gadea, Carlota Ontangas UPF · FORESEE project |
Software as a Medical Device A three-part session on regulatory compliance for medical software. Laura Becerra — Learning about Regulatory and QMS, the FORESEE case. Mar Gadea — Traceability matrix and Risk management assessment, with practical examples from the FORESEE project. Carlota Ontangas — Software Requirements Specification and Software Architecture Design, applying ISO 62304 across embedded software and clinical web applications. |
|
13 June 2025 15:00 · Room 52.219 Stefano Martina |
Quantum Computing and Quantum Machine Learning: recent advances The seminar will introduce the fundamentals of Quantum Computing, deep Quantum Machine Learning (QML) models and Quantum Neural Networks, and recent works from our group, including quantum diffusion models. |
|
29 May 2025 16:30 · Room 55.309 Simone Tassani, Carla Leal, Francis Chemorion UPF |
Data Driven Analysis A three-part session. Simone Tassani — Gait variability in osteoarthritic patients using dynamic time warping as treatment identifier. Carla Leal — Detailed cardiac functional and electrical phenotyping in propionic acidemia. Francis Chemorion — Disease-specific and shared metabolic signatures in atherosclerosis, knee osteoarthritis, and intervertebral disc degeneration: a population-based study. |
|
13 May 2025 11:30 · Room 55.309 Alin Tomescu Head of Cryptography, Aptos Labs |
Keyless blockchain accounts from zero-knowledge proofs Blockchain accounts are secured via secret keys, which has led to billions of dollars in lost or stolen assets. We propose a Keyless approach that requires only a traditional Web2 account (e.g., Google or Apple) and a privacy guardian that helps conceal your Web2 identity on-chain. Our method co-opts existing digital signatures from Web2 providers as transaction signatures and maintains privacy by encapsulating these signatures in a zero-knowledge proof. Keyless accounts have been deployed on the Aptos Labs blockchain for over a year. |
|
1 April 2025 12:30 · Room 52.019 Carlos Castillo, Manuel Portela, Martín Rocamora, Xavier Serra UPF |
Towards building Open and Fair AI: considerations about Data Governance, Open-Source models, and Human Oversight As artificial intelligence continues to shape industries and societies, the need for open, transparent, and fair systems becomes increasingly critical. This seminar explores key principles and practices that can guide the development of AI technologies that prioritize fairness, openness, and accountability. Carlos Castillo — Risks of AI-based decision support. Manuel Portela — Data governance and its importance for AI. Martín Rocamora and Xavier Serra — Open source AI models in music. |
|
26 March 2025 16:30 · Room 55.309 Jérôme Noailly, Aylin Kadkhodamanesh, Pablo Orons Vazquez UPF · Biomechanics and Mechanobiology team |
Biomechanics Deep Learning for Biomedical Research A three-part session. Jérôme Noailly — Innerspace: a computational journey through squeezing tissues and reacting cells, on modelling multifactorial processes in musculoskeletal joint degeneration. Aylin Kadkhodamanesh — Multiscale modeling of the Intervertebral disc to investigate the relationship between local degenerative processes and systemic low-grade inflammation. Pablo Orons Vazquez — SimOSpine: Navigating clinical needs and securing funding for success. |
|
26 February 2025 16:30 · Room 55.309 Miguel Angel González, Sikha O K, José Carlos Reyes UPF · SIMBIOsys |
Deep Learning for Biomedical Research A three-part session. Miguel Angel Gonzalez Ballester — Rage against the machine (learning): looking back at the work within SIMBIOsys and how deep learning has had a transformative effect. Sikha O K — From Black Box to Glass Box: Building Trust in AI with Explainability and Uncertainty. José Carlos Reyes Hernández — Towards Reliable Infant Face Detection and Landmark Estimation. |
|
28 January 2025 16:30 · Room 55.309 Bart Bijnens, Ainhoa Aguado, Carlos Albors UPF |
Genetic Cardiac Diseases, Platforms and In-silico Hemodynamics A three-part session. Bart Bijnens — Genetic Cardiac Diseases in the Plain People as compared to the English: why is it worth studying, and what can we learn from it. Ainhoa Aguado — Rocket Platform uncovered: the truth behind our platform. Carlos Albors — Impact of occluder device configurations in in-silico left atrial hemodynamics for the analysis of device-related thrombus. |
2024
|
17 December 2024 10:30 · Room Auditorium Núria Sebastián SAP Research Group, UPF Center for Brain and Cognition |
Building intelligent machines? Human intelligence at birth By examining the cognitive capabilities of newborns — such as basic perception, learning, and problem-solving — researchers seek to uncover foundational principles that could inform the creation of more adaptable and autonomous machines. Understanding how human intelligence unfolds may offer important parallels for designing intelligent machines capable of learning, adapting, and reasoning in ways that resemble human cognitive development. |
|
13 December 2024 10:00 · Room 55.410 Crescent Jicol Lecturer in Computer Science, University of Bath |
Decoding the User Experience: Towards a new generation of Personalized and Adaptive Virtual Reality Experiences VR head-mounted displays have received considerable interest from industry and academia alike. However, user-centred factors such as emotions and individual differences may outweigh technical features in shaping the VR experience. In this talk, I will share research supporting the importance of user factors and work contributing towards a vision for the future of VR: Personalized and Adaptive Experiences (PAEs), leveraging pre-experience adaptation to user traits and real-time adjustments based on user states. |
|
11 December 2024 10:00 · Room 52.323 Jacques Huyghe Bernal Chair for Biomedical Engineering |
Piezoelectricity signals connective tissue stress to voltage-gated ion channels Strain-generated potentials (SGPs) are known to activate mechanosensitive ion gating, previously attributed to electrokinetic phenomena such as streaming potentials. Here we provide evidence to support the potential contribution of piezoelectricity in a linear and reversible manner with magnitudes measured within human intervertebral discs. We present a model to demonstrate the effect of piezoelectricity of a hydrated extracellular space on the voltage-gating of ion channels. |
|
13 November 2024 11:00 · Room 55.309 Peter Grünwald CWI Amsterdam · Leiden University |
E is the new P E-values are an alternative to p-values that effortlessly deal with optional continuation: with e-value based tests and the corresponding anytime valid (AV) confidence intervals, one can always gather additional data, while keeping statistically valid conclusions. In this talk I will introduce e-values, e-processes and AV confidence intervals, and discuss in detail the relation to Bayesian approaches, which are in some ways very similar but in other ways quite different. |
|
15 July 2024 09:30 · Room Auditorium Zina Jarrahi Cinker Director General of MATTER · Chief Creator of PUZZLE X |
Open Conversation with Zina Jarrahi Cinker An open conversation with Dr. Zina Jarrahi Cinker, a recognized Frontier Tech strategist, condensed matter physicist and Deep Science advocate. She serves as the Director General of MATTER, an international think tank of 30 country chapters, and Chief Creator of PUZZLE X, the leading international event for Frontier Tech for the future. |
|
8 July 2024 14:30 · Room 55.309 Daniel Rueckert Alexander von Humboldt Professor for AI in Medicine, TU Munich · Imperial College London |
AI and the future of radiology Artificial Intelligence is changing many fields across science and our society. This talk will discuss how AI is changing medicine and healthcare, particularly in radiology. I will focus on how AI can support the acquisition of medical images and image analysis and interpretation, enabling the early detection of diseases and improved personalised diagnosis, with examples from neuro and cardiovascular MR imaging, and how AI solutions can be privacy-preserving while providing trustworthy and explainable solutions for clinicians. |
|
17 June 2024 11:00 · Room 55.410 Shlomi Laufer Assistant Professor, Technion – Israel Institute of Technology |
Enhancing Clinical Skill Assessment Through Computer Vision Medical education traditionally employs the apprenticeship model, which necessitates close follow-up and typically provides extremely subjective feedback. In this presentation, I will explore how advancements in computer vision can be utilized to create more informative assessments and feedback on surgical skills, including the roles of both surgeons and anesthesiologists, and how introducing cameras into the operating room provides a new avenue for analyzing surgical workflows. |
|
2 April 2024 12:30 · Room 55.410 Jim Goodell Editor, The Learning Engineering Toolkit · Chair, IEEE Learning Technology Standards Committee |
A Foundational Guide to Learning Engineering Learning engineering is a process and practice that applies the learning sciences using human-centered engineering design methodologies and data-informed decision-making to support learners and their development. This session is designed to expose foundational concepts and interconnections across various domains of learning engineering as background knowledge for future learning. |
|
19 March 2024 12:30 · Room 52.S29 Ágata Lapedriza |
Emotion perception: Challenges and use cases One of the challenges Human-Centered AI systems face is understanding human behavior and emotions considering the context in which they take place. Current computer vision approaches for recognizing human emotions usually focus on facial movements and often ignore the context. In this presentation, I will talk about our work on emotion perception and discuss the challenges we face when we attempt to create automatic models to perceive emotions, along with possible use cases. |
|
5 March 2024 12:30 · Room 52.219 Rubén Bonet Fractus |
Fractus, the Deep Tech pioneer In its academic origins at UPC Barcelona Tech, Fractus emerged as a pioneer in antenna technologies. Extensive R&D launched Fractus' technology to global recognition, accumulating adoption by a majority of mobile phone OEMs and network infrastructure operators. The company underwent a significant transformation, shifting its business model from product-centric to a licensing approach — offering valuable lessons for PhD students navigating their own paths of research and development. |
|
19 February 2024 15:00 · Room 55.309 Emma Robinson King's College London |
Computational methods for brain imaging analysis Dr Robinson's research focuses on the development of computational methods for brain imaging analysis, covering a wide range of image processing and machine learning topics. Her software for cortical surface registration (Multimodal Surface Matching, MSM) has been central to the Human Connectome Project's multi-modal parcellation of the human cortex. Current research interests focus on applying advanced machine learning, particularly Deep Learning, to diverse datasets combining multi-modality imaging data with genetic samples. |
|
19 February 2024 15:00 · Room 55.309 Silvia Budday |
Exploring brain mechanics Brain tissue is arguably the most complex and compliant tissue in the human body. Increasing evidence confirms that mechanics plays a critical role in modulating brain function and dysfunction. Computational models based on nonlinear continuum mechanics can help understand the basic processes in the brain during development, injury, and disease. This presentation will show how we develop computational models that capture both biological processes on the cellular scale and macroscopic loading and pathologies. |
2023
|
28 November 2023 15:30 · Room 55.309 Antonio Bonafonte, Guillermo Cambara Ruiz, Elena Sokolova Amazon |
Machine learning that powers Amazon Alexa In this presentation you will learn how machine learning is applied in industrial settings at scale. We will explain how Alexa works and what ML algorithms enable you to get a reply in less than a second. We will also talk about opportunities for students in Amazon and the hiring process for internships. |
|
24 November 2023 15:30 · Room 55.309 Zerihun G. Workineh, Morteza Rasouligandomani UPF · SIMBIOsys Biomechanics and Mechanobiology team |
Modelling in Biomechanics A new edition of BCN-MedTech Seminars, with dissemination and possible collaborations within our environment as the basis. This first session is about Modelling in Biomechanics, with two speakers from the Biomechanics and Mechanobiology team of the SIMBIOsys research group. Gemma Piella opens with a short introduction to the BCN MedTech group and its research. |
|
27 October 2023 12:00 · Room 53.309 Gemma De les Coves Associate Professor in Theoretical Physics, University of Innsbruck |
Universality in physics, computer science and beyond Certain spin models are universal, meaning that they contain all other models when seen in the appropriate light. What is the relation between universal spin models, universal Turing machines and notions of universality in neural networks? And what is their relation to forms of unreachability such as uncomputability or undecidability? In this talk, I will explore these questions in light of our recent framework for universality and share some philosophical perspectives on these structures. |
|
25 October 2023 11:30 · Room 55.309 Nicolás Escudero CEO, TRBL Services · European patent attorney |
Computer-based inventions: how to patent in the fields of software and artificial intelligence What is the best strategy to protect our research results? Can we patent software? And algorithms? Dr Nicolàs Escudero, an experienced European patent agent, shares his experience on how software is interpreted in patents, how to patent software, whether artificial intelligence can be patented, where to patent SW/AI, and his recommendations and conclusions. |
|
19 October 2023 12:30 · Room 52.019 Researchers from Kunstuniversität Graz (KUG) Institute for Ethnomusicology, Kunstuniversität Graz |
Computational Methods in Ethnomusicology: Case Studies from Kunstuniversität Graz Ethnomusicologists have always been prone to the use of technology for their research. This seminar presents four ethnomusicological research projects relying on computational methods: tango danceability analysis via motion capture, Balinese improvisatory solo drumming, the KUG Dastgāhi Corpus of Iranian dastgāhi music, and the study of perceptions of similarity in linear modal traditions. Each case study focuses on the research questions that motivated the use of computational tools, and reflects on the benefits and limitations in their implementation. |
|
2 October 2023 16:00 · Room 51.100 Roger Assaker CEO, MDSIM |
From Aerospace & Automotive to In-Silico Medicine The talk covers the entrepreneurial journey using innovative Computer Modeling & Simulations (CM&S) technology and Artificial Intelligence to solve a market need and create value, illustrated by two cases: e-Xstream engineering, which became the leader in advanced material modeling technology, and MDsim, a cross-industry innovation using CM&S and AI to improve the quality of life of patients suffering from spine degeneration and deformity. |
|
21 September 2023 16:00 · Room 55.309 Aleksei Tiulpin Assistant Professor, University of Oulu (IMEDS group) |
Osteoarthritis progression modeling and what we have been missing Dr. Tiulpin will discuss his past and recent research on modeling the progression of knee osteoarthritis using different modalities and methodological approaches, culminating with his lab's recent work on active surveillance. |
|
5 July 2023 10:30 · Room 52.119 John J. Costi Flinders University |
A multiaxial and multiscale assessment of lower back injury mechanisms Chronic low back pain is a crippling condition that affects quality of life and is a significant burden on the health care system. This presentation discusses a program of research on the multiscale assessment of the effects of disc degeneration and herniation. A sophisticated multiaxial hexapod robot was developed for simulating lifting loads on cadaver lumbar segments, together with the measurement of internal disc deformations for determining a lifting injury threshold. |
|
15 June 2023 15:00 · Room 51.100 Adrian Galdran Marie Skłodowska-Curie Research Fellow, UPF · Australian Institute of Machine Learning |
A Tutorial on Model Calibration for Neural Nets A calibrated machine learning model produces probabilistic predictions that are well-aligned with real probabilities: it tends to be more certain when it is correct. The goal of this talk is to introduce the main ideas and methods of modern model calibration: its formal definition, how we can measure it, and how we can improve it. The talk includes a short hands-on code session: we will train a skin lesion classifier in pytorch, measure its calibration, and learn a post-processing transform to improve it. |
|
24 February 2023 14:30 · Room 52.123 Schahram Dustdar Full Professor of Computer Science, TU Wien |
Edge Intelligence — Research Opportunities in the Distributed Computing Continuum As humans, Internet of Things, software services and AI continue to become the entangled fabric of distributed systems, systems engineers and researchers are facing novel challenges. In this talk, we analyze the role of AI in the context of IoT, Edge, and Cloud in the co-evolution of distributed systems for the new decade, identify challenges and discuss a roadmap to bring intelligence to the edge. |