EDIC Open House

Welcome!

Our Open House for admitted EDIC fellowship students will run for two days, from Tuesday, March 18 to Wednesday, March 19, 2025.

We planned a large number of short talks and group events, as well as plenty of opportunities for one-on-one meetings with IC faculty and students. These meetings are a great opportunity for attendees to learn more about the research we do, as well as ask all the questions about PhD student life at EPFL and in Lausanne.

We are looking forward to meeting you!

Thomas Bourgeat and Nicolas Flammarion, 2025 EDIC Open House Chairs

Schedule 2025

Day 1 – Tuesday, March 18
TimeProgramRoom
08:45-09:15Registration & welcome coffeeBC 430
09:20-12:30Welcome presentation & Lightning talks in 2 parallel tracksBC 420/410
12:30-14:00Standing lunch break & Poster SessionBC Hall
14:00-18:00One-on-one meetings
14:00-17:00Coffee time in the lab
18:00-19:00EPIC presentation & Campus tourINM 065
20:00 onDinner in town
Day 2 – Wednesday, March 19
TimeProgramRoom
09:00-12:00One-on-one meetings & Coffee time in the labs
12:00-12:20Closing remarksBC 420
12:30-13:45Lunch break
14:00-18:00One-on-one meetings & Coffee time in the labs
18:00 onApero featuring board games, EPIC, and current PhDsINM 065

Lightning talks
Parallel track I – BC 420
TimeTitle
09:20-10:10Welcome & Open House program overview – Welcome by PhD Reps
10:10-10:25Multi-modal and Generative AI for Biomedical Discovery (Maria Brbić)
10:25-10:40AI for health: from theory to applications (Dorina Thanou)
10:40-10:55Virtual Cells and Digital Twins: AI in Personalized Medicine (Charlotte Bunne)
10:55-11:20Break
11:25-11:40Deep Shape Modeling (Pascal Fua)
11:40-11:55FlexTok: Resampling Images into 1D Token Sequences of Flexible Length (Roman Bachmann)
11:55-12:10Language Models and the Brain: Alignment, Specialization, and Cognitive Parallels (Badr AlKhamissi)
12:10-12:25A phase transition between positional and semantic learning in a solvable model of dot-product attention (Freya Behrens)
12:25-12:40Physics-Informed Graph Neural Networks for Dynamical Systems (Olga Fink)
Parallel track II – BC 410
TimeTitle
10:10-10:25Making System Performance Programmable (George Candea)
10:25-10:40QuePaxa: Escaping the Tyranny of Timeouts in Consensus (Bryan Ford)
10:40-10:55Data-intensive Systems on Modern Hardware (Hamish Nicholson)
10:55-11:20Break
11:25-11:40E-Vote Your Conscience: Perceptions of Coercion and Vote Buying, and the Usability of Fake Credentials in Online Voting (Louis-Henri Merino)
11:40-11:55Pushing the Limits of Density in AI (Simla Harma)
11:55-12:10From coding theory to delegating computation (Giacomo Fenzi)
12:10-12:25Quantum Complexity (Thomas Vidick)
12:25-12:40Theory inspired by practice: quantum@EDIC (Yihui Quek, New IC/SB Faculty)

Poster list
TitleLaboratory
AI-powered virtual tissues from spatial proteomics for clinical diagnostics and biomedical discoveryArtificial Intelligence in Molecular Medicine (AIMM)
Test-Time View Selection for Multi-Modal Decision MakingArtificial Intelligence in Molecular Medicine (AIMM)
Semiparametric Triple Difference EstimatorsChair of Business Analystics (BAN)
Deep Geometric Learning for Engineering DesignCVLab
Learning Structured 3D Shape Representations for Design and OptimizationCVLab
Dalton: Learned Partitioning for Distributed Data StreamsData-Intensive Applications and Systems (DIAS)
Declarative Concurrent Data StructuresData-Intensive Applications and Systems (DIAS)
Efficient scheduling for multiple database systems on shared hardwareData-Intensive Applications and Systems (DIAS)
HPCache: Memory-Efficient OLAP through Proportional CachingData-Intensive Applications and Systems (DIAS)
Proteus: taming heterogeneity through virtualization & JIT adaptivityData-Intensive Applications and Systems (DIAS)
Adaptive Datalog / Recursive Query OptimizationData-Intensive Applications and Systems (DIAS) / Programming Methods Lab (LAMP)
E-Vote Your Conscience: Perceptions of Coercion and Vote Buying, and the Usability of Fake Credentials in Online VotingDecentralized Distributed Systems Laboratory (DEDIS)
QuePaxa: Escaping the Tyranny of Timeouts in ConsensusDecentralized Distributed Systems Laboratory (DEDIS)
Performance Interfaces For Hardware AcceleratorsDependable Systems Lab (DSLAB)
The Case of Energy ClarityDependable Systems Lab (DSLAB)
Transparent Multicore Scaling of Single-threaded Network FunctionsDependable Systems Lab (DSLAB)
Practical Verification of System-Software Components Written in Standard CDependable Systems Lab (DSLAB) / Systems and Formalisms Lab (SYSTEMF)
InkSight: Offline-to-Online Handwriting Conversion by Learning to Read and WriteGoogle DeepMind
SURGEON: Performant, Flexible, and Accurate Re-Hosting via TransplantationHexHive
type++: Prohibiting Type Confusion Through Inline Type InformationHexHive
Finding General HierarchiesInformation and Network Dynamics (INDY)
This Too Shall Pass: Removing Stale Observations in Dynamic Bayesian OptimizationInformation and Network Dynamics 2 (INDY2)
Fast Proxy Experiment Design for Causal Effect IdentificationInformation and Network Dynamics 2 (INDY2) / Chair of Business Analystics (BAN)
From Markov to Laplace: How Mamba In-Context Learns Markov ChainsInformation in Networked Systems (LINX)
Enhancing Multilingual LLM Pretraining with Model-Based Data SelectionMachine Learning and Optimization Laboratory (MLO)
Large (Vision) Language Models are Unsupervised In-Context LearnersMachine Learning for Biomedicine (MLBIO)
Tissue Reassembly Using Generative AIMachine Learning for Biomedicine (MLBIO)
Deep learning for measuring behaviorMathis Group
Skill learning and modeling sensorimotor circuitsMathis Group
IIBLAST: Accelerating Commercial FPGA CompilationParallel Systems Architecture Lab (PARSA)
Rack-Scale Servers for the Post-Moore EraParallel Systems Architecture Lab (PARSA)
Rebooting virtual memory with MidgardParallel Systems Architecture Lab (PARSA)
The Interplay between​ Quantization and Sparsity​Parallel Systems Architecture Lab (PARSA)
Fast, Flexible, and Practical Kernel ExtensionsRobust Scalable Systems Software Lab (RS3Lab)
Handling Latency in Tiered MemoryRobust Scalable Systems Software Lab (RS3Lab)
Monarch: A Fuzzing Framework for Distributed File SystemsRobust Scalable Systems Software Lab (RS3Lab)
Accelerating MoE Model Inference with Expert ShardingScalable Computing Systems (SaCS)
Revisiting Ensembling in One-Shot Federated LearningScalable Computing Systems (SaCS)
A Phase Transition between Positional and Semantic Learning in a Solvable Model of Dot-Product AttentionStatistical Physics of Computations (SPOC)
Bayes-optimal learning of an extensive-width neural network from quadratically many samplesStatistical Physics of Computations (SPOC)
Diagrammatic notations: a wishlistSystems and Formalisms Lab (SYSTEMF)
A Language for Microarchitectural Security EvaluationVerification and Computer Architecture (VCA)
DemoGeometric Computing Laboratory (GCM)

Coffee time in the labs
Tuesday March 18
TimeGroupRoom
14:00-15:00Geometric Computing LaboratoryINM014
14:00-15:30Mathematical Data Science ChairINR 112
14:00-16:00Coffee in the AIMM labBC 247
15:00-16:00Coffee at MLBIO labINJ 321
16:00-17:00Theory of Machine Learning LabINR 120
PLFM tea hourINN 326
17:00-18:00Lab for intelligent Global Health and Humanitarian Response TechnologiesBC 410
Wednesday March 19
TimeGroupRoom
9:00-10:00HexHiveBC 159
10:00-11:00Coffee & Croissant at PARSAINJ 213
11:00-12:00Coffee/Snacks with RS3LabINN 212
14:00-15:00Swiss AI Initiative Coffee & SnacksELE 117
15:00-16:00Coffee in the DIAS labBC 229
Coffee in the VILABINJ 211
16:00-17:00Coffee in the SPOC labBSP 507
Crypto+Quantum coffee in the COMPSEC labBC 242

Experience Switzerland

  • MySwitzerland.com. In-depth, multilingual Switzerland Tourism website (VR experience, videos, maps, etc.).
  • Virtual tour of the Canton of Vaud. Navigate this interactive portal to explore some of the most beautiful sites in our region from your sofa.
  • Lake Geneva Region

EDIC Useful links