Event owner lens
Use this record to benchmark whether a comparable event makes the work after the room visible. The score is 35/100, so the next move is to benchmark the weakest pillars before repeating the format.
broadcast-heavy conference agenda analysis
Iclr 2026
This broadcast-heavy conference in Academic / Research / Science shows 70 visible agenda rows from iclr.cc and scores 35/100: a thin but inspectable design signal. The clearest public signals sit in Learning Transfer and Problem Specificity; the main limits are Follow Through and Network Design. Visible mechanisms include Participant work, Feedback, Network design, and Learning transfer. The public record does not show follow-up or tracking, so the score should be read as design intent rather than durable impact. A practical reading: For a reader, this is a comparison record more than a model to copy: it reads as a broadcast-heavy conference, with the strongest visible signal in learning transfer and problem specificity and the biggest open question around follow through and network design. The practical test is whether the published agenda connects the room to post-event continuation and evidence. This page is an original public-evidence analysis, not a copy of the source agenda or an endorsement of the event. The score places the visible agenda in the thin outcome architecture band. The strongest visible pillars are Learning Transfer, Problem Specificity, and Evidence Maturity; the thinnest visible pillars are Follow Through, Network Design, and Participation Architecture. Visible mechanisms include Participant work, Feedback, Network design, Learning transfer, and Personalization. The extracted agenda preview includes 132 visible rows. The most common formats are Unknown, Presentation, and Poster Session; the most common inferred purposes are Unknown, Knowledge Transfer, and Showcase.
Eight-pillar fingerprint
Hover any pillar to see what it measures and, where it scored low, what the agenda is missing.
Participation Architecture?30
Participation Architecture - 30/100. Participant work, contribution, interaction, and alternatives to passive broadcast.Missing: Turn passive airtime into participant work: practice, sensemaking, decisions, critique, or artifact creation.
Follow Through?5
Follow Through - 5/100. Owners, dates, commitments, progress checks, and accountability after the room.Missing: Add named owners, dates, implementation checkpoints, and a visible post-event continuation path.
Problem Specificity?46
Problem Specificity - 46/100. A clear costly problem, objective, decision, or performance target.
Personalization?38
Personalization - 38/100. Role, path, goal, preparation, or connection tailoring for participants.
Network Design?16
Network Design - 16/100. Structured weak ties, bridge-building, mixers, and relationship persistence.Missing: Replace generic networking blocks with designed introductions, ask-offer exchanges, peer groups, or bridge-building rituals.
Learning Transfer?61
Learning Transfer - 61/100. Applied practice, feedback, workplace use, refreshers, and 30-90 day transfer.
Evidence Maturity?41
Evidence Maturity - 41/100. Baseline, comparison, follow-up, isolation, and attribution confidence.
Future-of-Work Fit?38
Future-of-Work Fit - 38/100. Value against time, hybrid reality, accessibility, AI, and meeting load.
Fix the gaps
Field-tested exercises matched to this agenda's weakest pillars, from the exercise library.
Follow Through (5/100)
Exercises that strengthen it: Mental Toughness Workshop • WorkshopBank · 15% Solutions • WorkshopBank · Network Patches
Network Design (16/100)
Exercises that strengthen it: Fun Trust-Building Group Game for Team Connection - Izzat... · Exciting Group Trust Activity for Building Confidence... · PFP: September 2005
Participation Architecture (30/100)
Exercises that strengthen it: Make A World · Awestruck 3 Minutes · Spectrum Mapping
Agenda Preview
The actual agenda we captured. Every block is classified by format and purpose. Open any block to see how we read it; the colored edge shows whether it is participant work, broadcast, logistics, or a showcase.
Room vs wrapper
5 percent of the 132 classified blocks put participants to work; the rest broadcast, show, or handle logistics. That mix is what drives the participation score.
Participant workBroadcastShowcaseLogistics
all eventBlog Track PostersPoster sessionShowcase+
Format · ShowcasePoster sessionPresenters display work; attendees browse and ask questions. Some interaction, not structured work.
Evidence basisOutcome inferred from formatInferred from format
all eventJournal Track PostersPoster sessionShowcase+
Format · ShowcasePoster sessionPresenters display work; attendees browse and ask questions. Some interaction, not structured work.
Evidence basisOutcome inferred from formatInferred from format
all eventWorkshopsWorkshopParticipant work+
Format · Participant workWorkshopParticipants work on a problem and produce something. The strongest signal of participation architecture.
Evidence basisParticipant work is implied by the formatInferred from format
all eventIn-context learning of representations can be explained by induction circuitsPresentationKnowledge transfer+
Format · BroadcastPresentationSpeakers present, the audience receives. Awareness only unless paired with practice or follow-up.
Evidence basisNo participant output visible from this rowRead from source, no work signal
all eventFrom U-Nets to DiTs: The Architectural Evolution of Text-to-Image Diffusion Models (2021 to 2025)PresentationKnowledge transfer+
Format · BroadcastPresentationSpeakers present, the audience receives. Awareness only unless paired with practice or follow-up.
Evidence basisNo participant output visible from this rowRead from source, no work signal
all eventAn Overview of Subliminal LearningPresentationKnowledge transfer+
Format · BroadcastPresentationSpeakers present, the audience receives. Awareness only unless paired with practice or follow-up.
Evidence basisNo participant output visible from this rowRead from source, no work signal
all eventFrom REINFORCE to Dr. GRPO: A Unified Perspective on LLM Post-TrainingTrainingSkill building+
Format · Participant workTrainingGuided skill building where participants practice. Counts as participant work and learning transfer.
Evidence basisParticipant work is implied by the formatInferred from format
all eventFrom Trajectories to Operators - A Unified Flow Map Perspective on Generative ModelingPresentationKnowledge transfer+
Format · BroadcastPresentationSpeakers present, the audience receives. Awareness only unless paired with practice or follow-up.
Evidence basisNo participant output visible from this rowRead from source, no work signal
all eventLanguage as a Window Into the Mind: How NLP and LLMs Advance Human SciencesPresentationKnowledge transfer+
Format · BroadcastPresentationSpeakers present, the audience receives. Awareness only unless paired with practice or follow-up.
Evidence basisNo participant output visible from this rowRead from source, no work signal
all eventChunkTabPFN: Training-free Long ContextTrainingSkill building+
Format · Participant workTrainingGuided skill building where participants practice. Counts as participant work and learning transfer.
Evidence basisParticipant work is implied by the formatInferred from format
all eventNavigating the Manifold - A Geometric Perspective on Diffusion-Based Inverse ProblemsPresentationKnowledge transfer+
Format · BroadcastPresentationSpeakers present, the audience receives. Awareness only unless paired with practice or follow-up.
Evidence basisNo participant output visible from this rowRead from source, no work signal
all eventHow To Open the Black Box: Modern Models for Mechanistic InterpretabilityPresentationKnowledge transfer+
Format · BroadcastPresentationSpeakers present, the audience receives. Awareness only unless paired with practice or follow-up.
Evidence basisNo participant output visible from this rowRead from source, no work signal
all eventIs the evidence in 'Language Models Learn to Mislead Humans via RLHF' valid?PresentationKnowledge transfer+
Format · BroadcastPresentationSpeakers present, the audience receives. Awareness only unless paired with practice or follow-up.
Evidence basisNo participant output visible from this rowRead from source, no work signal
all eventExtracting Model Precision from 20 LogprobsPresentationKnowledge transfer+
Format · BroadcastPresentationSpeakers present, the audience receives. Awareness only unless paired with practice or follow-up.
Evidence basisNo participant output visible from this rowRead from source, no work signal
all eventEffect of Parallel Environments and Rollout Steps in PPOPresentationKnowledge transfer+
Format · BroadcastPresentationSpeakers present, the audience receives. Awareness only unless paired with practice or follow-up.
Evidence basisNo participant output visible from this rowRead from source, no work signal
all eventWait, Do We Need to Wait? Revisiting Budget Forcing for Sequential Test-Time ScalingPresentationKnowledge transfer+
Format · BroadcastPresentationSpeakers present, the audience receives. Awareness only unless paired with practice or follow-up.
Evidence basisNo participant output visible from this rowRead from source, no work signal
all eventVisualizing LLM Latent Space Geometry Through Dimensionality ReductionPresentationKnowledge transfer+
Format · BroadcastPresentationSpeakers present, the audience receives. Awareness only unless paired with practice or follow-up.
Evidence basisNo participant output visible from this rowRead from source, no work signal
all eventUsing Graph Neural Networks in Reinforcement Learning: A Practical GuidePresentationKnowledge transfer+
Format · BroadcastPresentationSpeakers present, the audience receives. Awareness only unless paired with practice or follow-up.
Evidence basisNo participant output visible from this rowRead from source, no work signal
all eventTrade-offs in LLM Compute for Reasoning-Intensive Information RetrievalPresentationKnowledge transfer+
Format · BroadcastPresentationSpeakers present, the audience receives. Awareness only unless paired with practice or follow-up.
Evidence basisNo participant output visible from this rowRead from source, no work signal
all eventArtistic Style and the Play of Neural Style RepresentationsPresentationKnowledge transfer+
Format · BroadcastPresentationSpeakers present, the audience receives. Awareness only unless paired with practice or follow-up.
Evidence basisNo participant output visible from this rowRead from source, no work signal
3:15 PM - 5:45 PMApr 24, Pavilion 3PresentationKnowledge transfer+
Format · BroadcastPresentationSpeakers present, the audience receives. Awareness only unless paired with practice or follow-up.
Evidence basisNo participant output visible from this rowRead from source, no work signal
all eventThe effect of feature resolution on embedding dimensionPresentationKnowledge transfer+
Format · BroadcastPresentationSpeakers present, the audience receives. Awareness only unless paired with practice or follow-up.
Evidence basisNo participant output visible from this rowRead from source, no work signal
10:30 AM - 1:00 PMApr 25, Pavilion 3PresentationKnowledge transfer+
Format · BroadcastPresentationSpeakers present, the audience receives. Awareness only unless paired with practice or follow-up.
Evidence basisNo participant output visible from this rowRead from source, no work signal
all eventDivide, Conquer, and Standardize - A Recursive Architecture for Multi-Agent Systems (MAS)PresentationKnowledge transfer+
Format · BroadcastPresentationSpeakers present, the audience receives. Awareness only unless paired with practice or follow-up.
Evidence basisNo participant output visible from this rowRead from source, no work signal
all eventWhat (and What Not) are Calibrated Probabilities Actually Useful for?PresentationKnowledge transfer+
Format · BroadcastPresentationSpeakers present, the audience receives. Awareness only unless paired with practice or follow-up.
Evidence basisNo participant output visible from this rowRead from source, no work signal
all eventThe Layered Ontology of Models, Resolving the Epistemological Crisis of AIPresentationKnowledge transfer+
Format · BroadcastPresentationSpeakers present, the audience receives. Awareness only unless paired with practice or follow-up.
Evidence basisNo participant output visible from this rowRead from source, no work signal
all eventThe 99% Success Paradox: When Near-Perfect Retrieval Equals Random SelectionPresentationKnowledge transfer+
Format · BroadcastPresentationSpeakers present, the audience receives. Awareness only unless paired with practice or follow-up.
Evidence basisNo participant output visible from this rowRead from source, no work signal
all eventFaster SVD via Accelerated Newton-Schulz IterationPresentationKnowledge transfer+
Format · BroadcastPresentationSpeakers present, the audience receives. Awareness only unless paired with practice or follow-up.
Evidence basisNo participant output visible from this rowRead from source, no work signal
10:30 AM - 1:00 PMApr 23, Pavilion 3PresentationKnowledge transfer+
Format · BroadcastPresentationSpeakers present, the audience receives. Awareness only unless paired with practice or follow-up.
Evidence basisNo participant output visible from this rowRead from source, no work signal
all eventWhere’s the Chicken? Unpacking Spatial Awareness in Vision-Language ModelsPresentationKnowledge transfer+
Format · BroadcastPresentationSpeakers present, the audience receives. Awareness only unless paired with practice or follow-up.
Evidence basisNo participant output visible from this rowRead from source, no work signal
10:30 AM - 1:00 PMApr 25, Pavilion 4PresentationKnowledge transfer+
Format · BroadcastPresentationSpeakers present, the audience receives. Awareness only unless paired with practice or follow-up.
Evidence basisNo participant output visible from this rowRead from source, no work signal
all eventdLLM - Rethinking Generation Beyond Autoregressive ModelsPresentationKnowledge transfer+
Format · BroadcastPresentationSpeakers present, the audience receives. Awareness only unless paired with practice or follow-up.
Evidence basisNo participant output visible from this rowRead from source, no work signal
all eventRethinking the Diffusion Model from a Langevin PerspectivePresentationKnowledge transfer+
Format · BroadcastPresentationSpeakers present, the audience receives. Awareness only unless paired with practice or follow-up.
Evidence basisNo participant output visible from this rowRead from source, no work signal
all eventDissecting Non-Determinism in Large Language ModelsPresentationKnowledge transfer+
Format · BroadcastPresentationSpeakers present, the audience receives. Awareness only unless paired with practice or follow-up.
Evidence basisNo participant output visible from this rowRead from source, no work signal
10:30 AM - 1:00 PMApr 23, Pavilion 4PresentationKnowledge transfer+
Format · BroadcastPresentationSpeakers present, the audience receives. Awareness only unless paired with practice or follow-up.
Evidence basisNo participant output visible from this rowRead from source, no work signal
all eventThe human knowledge loophole in the 'bitter lesson' for LLMsPresentationKnowledge transfer+
Format · BroadcastPresentationSpeakers present, the audience receives. Awareness only unless paired with practice or follow-up.
Evidence basisNo participant output visible from this rowRead from source, no work signal
all eventTracing the Principles Behind Modern Diffusion ModelsPresentationKnowledge transfer+
Format · BroadcastPresentationSpeakers present, the audience receives. Awareness only unless paired with practice or follow-up.
Evidence basisNo participant output visible from this rowRead from source, no work signal
all eventDiffusion as Infinite HVAEs: Do Diffusion Models Generalize Better than Deep VAEs?PresentationKnowledge transfer+
Format · BroadcastPresentationSpeakers present, the audience receives. Awareness only unless paired with practice or follow-up.
Evidence basisNo participant output visible from this rowRead from source, no work signal
all eventThe Adversarial Conditioning Paradox: Why Attacked Inputs Are More Stable, Not LessPresentationKnowledge transfer+
Format · BroadcastPresentationSpeakers present, the audience receives. Awareness only unless paired with practice or follow-up.
Evidence basisNo participant output visible from this rowRead from source, no work signal
all eventEvaluating Machine Learned Inter-Atomic Potentials for a Practical Simulation WorkflowPresentationKnowledge transfer+
Format · BroadcastPresentationSpeakers present, the audience receives. Awareness only unless paired with practice or follow-up.
Evidence basisNo participant output visible from this rowRead from source, no work signal
all eventDynamic Parameter Reuse Augments Reasoning via Latent Chain of ThoughtPresentationKnowledge transfer+
Format · BroadcastPresentationSpeakers present, the audience receives. Awareness only unless paired with practice or follow-up.
Evidence basisNo participant output visible from this rowRead from source, no work signal
all eventDestruction is a General Strategy to Learn Generation; Diffusion's Strength is to Take it Seriously; Exploration is the FuturePresentationKnowledge transfer+
Format · BroadcastPresentationSpeakers present, the audience receives. Awareness only unless paired with practice or follow-up.
Evidence basisNo participant output visible from this rowRead from source, no work signal
10:30 AM - 1:00 PMApr 24, Pavilion 4PresentationKnowledge transfer+
Format · BroadcastPresentationSpeakers present, the audience receives. Awareness only unless paired with practice or follow-up.
Evidence basisNo participant output visible from this rowRead from source, no work signal
all eventDefining and quantifying compositional structurePresentationKnowledge transfer+
Format · BroadcastPresentationSpeakers present, the audience receives. Awareness only unless paired with practice or follow-up.
Evidence basisNo participant output visible from this rowRead from source, no work signal
all eventModel Misspecification in Simulation-Based Inference - Recent Advances and Open ChallengesPresentationKnowledge transfer+
Format · BroadcastPresentationSpeakers present, the audience receives. Awareness only unless paired with practice or follow-up.
Evidence basisNo participant output visible from this rowRead from source, no work signal
all eventLearning to Maximize Rewards via Reaching GoalsPresentationKnowledge transfer+
Format · BroadcastPresentationSpeakers present, the audience receives. Awareness only unless paired with practice or follow-up.
Evidence basisNo participant output visible from this rowRead from source, no work signal
all eventJustRL: Scaling a 1.5B LLM with a Simple RL RecipePresentationKnowledge transfer+
Format · BroadcastPresentationSpeakers present, the audience receives. Awareness only unless paired with practice or follow-up.
Evidence basisNo participant output visible from this rowRead from source, no work signal
all eventBudget Alignment: Making Models Reason in the User's LanguagePresentationKnowledge transfer+
Format · BroadcastPresentationSpeakers present, the audience receives. Awareness only unless paired with practice or follow-up.
Evidence basisNo participant output visible from this rowRead from source, no work signal
all eventUnigramLM: An Attempt at Writing The Missing ManualPresentationKnowledge transfer+
Format · BroadcastPresentationSpeakers present, the audience receives. Awareness only unless paired with practice or follow-up.
Evidence basisNo participant output visible from this rowRead from source, no work signal
all eventHeuristic-Based Ideation for Guiding LLMs Toward Structured CreativityPresentationKnowledge transfer+
Format · BroadcastPresentationSpeakers present, the audience receives. Awareness only unless paired with practice or follow-up.
Evidence basisNo participant output visible from this rowRead from source, no work signal
all eventRevisiting the NetHack Learning EnvironmentPresentationKnowledge transfer+
Format · BroadcastPresentationSpeakers present, the audience receives. Awareness only unless paired with practice or follow-up.
Evidence basisNo participant output visible from this rowRead from source, no work signal
all eventContent Promotion as a Strategic Game: How to Design Agentic Publishers for the Evolving Search Ecosystem in the GenAI Era?PresentationKnowledge transfer+
Format · BroadcastPresentationSpeakers present, the audience receives. Awareness only unless paired with practice or follow-up.
Evidence basisNo participant output visible from this rowRead from source, no work signal
all eventUnderstanding and Fixing Bottlenecks in State Space Models: What Recency and Over-Smoothing Tell UsPresentationKnowledge transfer+
Format · BroadcastPresentationSpeakers present, the audience receives. Awareness only unless paired with practice or follow-up.
Evidence basisNo participant output visible from this rowRead from source, no work signal
all eventSquare Peg, Round Hole: Plugging Non-Sequential Data into Sequential Language ModelsPresentationKnowledge transfer+
Format · BroadcastPresentationSpeakers present, the audience receives. Awareness only unless paired with practice or follow-up.
Evidence basisNo participant output visible from this rowRead from source, no work signal
all eventReady For General Agents? Let's test it.PresentationKnowledge transfer+
Format · BroadcastPresentationSpeakers present, the audience receives. Awareness only unless paired with practice or follow-up.
Evidence basisNo participant output visible from this rowRead from source, no work signal
all eventDon't Look Up (Every Token): Escaping Quadratic Complexity via Geometric Patterns and AlgorithmsPresentationKnowledge transfer+
Format · BroadcastPresentationSpeakers present, the audience receives. Awareness only unless paired with practice or follow-up.
Evidence basisNo participant output visible from this rowRead from source, no work signal
all eventMisalignments and RL Failure Modes in the Early Stage of SuperintelligencePresentationKnowledge transfer+
Format · BroadcastPresentationSpeakers present, the audience receives. Awareness only unless paired with practice or follow-up.
Evidence basisNo participant output visible from this rowRead from source, no work signal
all eventComputer Use Survey - A Visual Survey of Computer Use AgentsPresentationKnowledge transfer+
Format · BroadcastPresentationSpeakers present, the audience receives. Awareness only unless paired with practice or follow-up.
Evidence basisNo participant output visible from this rowRead from source, no work signal
all eventProbabilistic Circuits for Uncertainty QuantificationPresentationKnowledge transfer+
Format · BroadcastPresentationSpeakers present, the audience receives. Awareness only unless paired with practice or follow-up.
Evidence basisNo participant output visible from this rowRead from source, no work signal
all eventLoneliness as a Case Study for Social Reward MisalignmentPresentationKnowledge transfer+
Format · BroadcastPresentationSpeakers present, the audience receives. Awareness only unless paired with practice or follow-up.
Evidence basisNo participant output visible from this rowRead from source, no work signal
all eventWhy AI Evaluations Need Error BarsPresentationKnowledge transfer+
Format · BroadcastPresentationSpeakers present, the audience receives. Awareness only unless paired with practice or follow-up.
Evidence basisNo participant output visible from this rowRead from source, no work signal
all eventAI Fundamentals: Valuing AI Agents & Data AssetsPresentationKnowledge transfer+
Format · BroadcastPresentationSpeakers present, the audience receives. Awareness only unless paired with practice or follow-up.
Evidence basisNo participant output visible from this rowRead from source, no work signal
all eventBlog Track PostersPoster SessionShowcase+
Format · ShowcasePoster SessionPresenters display work; attendees browse and ask questions. Some interaction, not structured work.
Evidence basisMediumRead from source
all eventJournal Track PostersPoster SessionShowcase+
Format · ShowcasePoster SessionPresenters display work; attendees browse and ask questions. Some interaction, not structured work.
Evidence basisMediumRead from source
all eventWorkshopsWorkshopCo Creation+
Format · Participant workWorkshopParticipants work on a problem and produce something. The strongest signal of participation architecture.
Evidence basisMediumRead from source
all eventEvolution of Flash AttentionUnknownUnknown+
Format · BroadcastUnknownFormat not classified from the source; treated as a broadcast block by default.
Evidence basisLowRead from source
all eventView full detailsUnknownUnknown+
Format · BroadcastUnknownFormat not classified from the source; treated as a broadcast block by default.
Evidence basisLowRead from source
all eventIn-context learning of representations can be explained by induction circuitsPresentationKnowledge Transfer+
Format · BroadcastPresentationSpeakers present, the audience receives. Awareness only unless paired with practice or follow-up.
Evidence basisMediumRead from source
all eventFrom U-Nets to DiTs: The Architectural Evolution of Text-to-Image Diffusion Models (2021 to 2025)UnknownUnknown+
Format · BroadcastUnknownFormat not classified from the source; treated as a broadcast block by default.
Evidence basisLowRead from source
all eventAn Overview of Subliminal LearningUnknownUnknown+
Format · BroadcastUnknownFormat not classified from the source; treated as a broadcast block by default.
Evidence basisLowRead from source
all eventFrom REINFORCE to Dr. GRPO: A Unified Perspective on LLM Post-TrainingTrainingSkill Building+
Format · Participant workTrainingGuided skill building where participants practice. Counts as participant work and learning transfer.
Evidence basisMediumRead from source
all eventFrom Trajectories to Operators - A Unified Flow Map Perspective on Generative ModelingUnknownUnknown+
Format · BroadcastUnknownFormat not classified from the source; treated as a broadcast block by default.
Evidence basisLowRead from source
all eventLanguage as a Window Into the Mind: How NLP and LLMs Advance Human SciencesUnknownUnknown+
Format · BroadcastUnknownFormat not classified from the source; treated as a broadcast block by default.
Evidence basisLowRead from source
all eventChunkTabPFN: Training-free Long ContextTrainingSkill Building+
Format · Participant workTrainingGuided skill building where participants practice. Counts as participant work and learning transfer.
Evidence basisMediumRead from source
all eventNavigating the Manifold - A Geometric Perspective on Diffusion-Based Inverse ProblemsUnknownUnknown+
Format · BroadcastUnknownFormat not classified from the source; treated as a broadcast block by default.
Evidence basisLowRead from source
all eventHow To Open the Black Box: Modern Models for Mechanistic InterpretabilityUnknownUnknown+
Format · BroadcastUnknownFormat not classified from the source; treated as a broadcast block by default.
Evidence basisLowRead from source
all eventIs the evidence in 'Language Models Learn to Mislead Humans via RLHF' valid?UnknownUnknown+
Format · BroadcastUnknownFormat not classified from the source; treated as a broadcast block by default.
Evidence basisLowRead from source
all eventExtracting Model Precision from 20 LogprobsUnknownUnknown+
Format · BroadcastUnknownFormat not classified from the source; treated as a broadcast block by default.
Evidence basisLowRead from source
all eventEffect of Parallel Environments and Rollout Steps in PPOUnknownUnknown+
Format · BroadcastUnknownFormat not classified from the source; treated as a broadcast block by default.
Evidence basisLowRead from source
all eventWait, Do We Need to Wait? Revisiting Budget Forcing for Sequential Test-Time ScalingUnknownUnknown+
Format · BroadcastUnknownFormat not classified from the source; treated as a broadcast block by default.
Evidence basisLowRead from source
all eventVisualizing LLM Latent Space Geometry Through Dimensionality ReductionUnknownUnknown+
Format · BroadcastUnknownFormat not classified from the source; treated as a broadcast block by default.
Evidence basisLowRead from source
all eventUsing Graph Neural Networks in Reinforcement Learning: A Practical GuideUnknownUnknown+
Format · BroadcastUnknownFormat not classified from the source; treated as a broadcast block by default.
Evidence basisLowRead from source
all eventTrade-offs in LLM Compute for Reasoning-Intensive Information RetrievalUnknownUnknown+
Format · BroadcastUnknownFormat not classified from the source; treated as a broadcast block by default.
Evidence basisLowRead from source
all eventArtistic Style and the Play of Neural Style RepresentationsPresentationKnowledge Transfer+
Format · BroadcastPresentationSpeakers present, the audience receives. Awareness only unless paired with practice or follow-up.
Evidence basisMediumRead from source
3:15 PM - 5:45 PMApr 24, Pavilion 3UnknownUnknown+
Format · BroadcastUnknownFormat not classified from the source; treated as a broadcast block by default.
Evidence basisMediumRead from source
all eventThe effect of feature resolution on embedding dimensionUnknownUnknown+
Format · BroadcastUnknownFormat not classified from the source; treated as a broadcast block by default.
Evidence basisLowRead from source
10:30 AM - 1:00 PMApr 25, Pavilion 3UnknownUnknown+
Format · BroadcastUnknownFormat not classified from the source; treated as a broadcast block by default.
Evidence basisMediumRead from source
all eventDivide, Conquer, and Standardize - A Recursive Architecture for Multi-Agent Systems (MAS)UnknownUnknown+
Format · BroadcastUnknownFormat not classified from the source; treated as a broadcast block by default.
Evidence basisLowRead from source
all eventWhat (and What Not) are Calibrated Probabilities Actually Useful for?UnknownUnknown+
Format · BroadcastUnknownFormat not classified from the source; treated as a broadcast block by default.
Evidence basisLowRead from source
all eventThe Layered Ontology of Models, Resolving the Epistemological Crisis of AIUnknownUnknown+
Format · BroadcastUnknownFormat not classified from the source; treated as a broadcast block by default.
Evidence basisLowRead from source
all eventThe 99% Success Paradox: When Near-Perfect Retrieval Equals Random SelectionUnknownUnknown+
Format · BroadcastUnknownFormat not classified from the source; treated as a broadcast block by default.
Evidence basisLowRead from source
all eventFaster SVD via Accelerated Newton-Schulz IterationUnknownUnknown+
Format · BroadcastUnknownFormat not classified from the source; treated as a broadcast block by default.
Evidence basisLowRead from source
10:30 AM - 1:00 PMApr 23, Pavilion 3UnknownUnknown+
Format · BroadcastUnknownFormat not classified from the source; treated as a broadcast block by default.
Evidence basisMediumRead from source
all eventWhere’s the Chicken? Unpacking Spatial Awareness in Vision-Language ModelsUnknownUnknown+
Format · BroadcastUnknownFormat not classified from the source; treated as a broadcast block by default.
Evidence basisLowRead from source
10:30 AM - 1:00 PMApr 25, Pavilion 4UnknownUnknown+
Format · BroadcastUnknownFormat not classified from the source; treated as a broadcast block by default.
Evidence basisMediumRead from source
all eventdLLM - Rethinking Generation Beyond Autoregressive ModelsUnknownUnknown+
Format · BroadcastUnknownFormat not classified from the source; treated as a broadcast block by default.
Evidence basisLowRead from source
all eventRethinking the Diffusion Model from a Langevin PerspectiveUnknownUnknown+
Format · BroadcastUnknownFormat not classified from the source; treated as a broadcast block by default.
Evidence basisLowRead from source
all eventDissecting Non-Determinism in Large Language ModelsUnknownUnknown+
Format · BroadcastUnknownFormat not classified from the source; treated as a broadcast block by default.
Evidence basisLowRead from source
10:30 AM - 1:00 PMApr 23, Pavilion 4UnknownUnknown+
Format · BroadcastUnknownFormat not classified from the source; treated as a broadcast block by default.
Evidence basisMediumRead from source
all eventThe human knowledge loophole in the 'bitter lesson' for LLMsUnknownUnknown+
Format · BroadcastUnknownFormat not classified from the source; treated as a broadcast block by default.
Evidence basisLowRead from source
all eventTracing the Principles Behind Modern Diffusion ModelsUnknownUnknown+
Format · BroadcastUnknownFormat not classified from the source; treated as a broadcast block by default.
Evidence basisLowRead from source
all eventDiffusion as Infinite HVAEs: Do Diffusion Models Generalize Better than Deep VAEs?UnknownUnknown+
Format · BroadcastUnknownFormat not classified from the source; treated as a broadcast block by default.
Evidence basisLowRead from source
all eventThe Adversarial Conditioning Paradox: Why Attacked Inputs Are More Stable, Not LessUnknownUnknown+
Format · BroadcastUnknownFormat not classified from the source; treated as a broadcast block by default.
Evidence basisLowRead from source
all eventPerformative Prediction made practicalUnknownUnknown+
Format · BroadcastUnknownFormat not classified from the source; treated as a broadcast block by default.
Evidence basisLowRead from source
all eventEvaluating Machine Learned Inter-Atomic Potentials for a Practical Simulation WorkflowUnknownUnknown+
Format · BroadcastUnknownFormat not classified from the source; treated as a broadcast block by default.
Evidence basisLowRead from source
all eventDynamic Parameter Reuse Augments Reasoning via Latent Chain of ThoughtUnknownUnknown+
Format · BroadcastUnknownFormat not classified from the source; treated as a broadcast block by default.
Evidence basisLowRead from source
all eventStandard language models often rely on massive parameter counts for their performance, utilizing each parameter only once per inference pass. This prompts consideration of recurrent structures, where models reuse parameters across sequential time, depth, or training progression to achieve improved performance and reduced training cost. We draw connections in the landscape of parameter reuse, from growing models via stacking to recurrent looping, and postulate that these architectural priors act as a form of Latent Chain of Thought (LCoT), allowing models to reason in a continuous state space. By shifting towards deeper and dynamic computation, grown and recurrent architectures offer a path toward improved reasoning in compact networks, ascending beyond scaling laws of standard architectures.TrainingSkill Building+
Format · Participant workTrainingGuided skill building where participants practice. Counts as participant work and learning transfer.
Evidence basisMediumRead from source
all eventDestruction is a General Strategy to Learn Generation; Diffusion's Strength is to Take it Seriously; Exploration is the FutureUnknownUnknown+
Format · BroadcastUnknownFormat not classified from the source; treated as a broadcast block by default.
Evidence basisLowRead from source
10:30 AM - 1:00 PMApr 24, Pavilion 4UnknownUnknown+
Format · BroadcastUnknownFormat not classified from the source; treated as a broadcast block by default.
Evidence basisMediumRead from source
all eventDefining and quantifying compositional structureUnknownUnknown+
Format · BroadcastUnknownFormat not classified from the source; treated as a broadcast block by default.
Evidence basisLowRead from source
all eventModel Misspecification in Simulation-Based Inference - Recent Advances and Open ChallengesUnknownUnknown+
Format · BroadcastUnknownFormat not classified from the source; treated as a broadcast block by default.
Evidence basisLowRead from source
all eventModel misspecification is a critical challenge in simulation-based inference (SBI), particularly in neural SBI methods that use simulated data to train flexible neural density estimators. These methods typically assume that simulators faithfully represent the true data-generating process, an assumption that is often violated in practice. Resulting discrepancies can make observed data effectively out-of-distribution relative to the simulations, leading to biased posterior distributions and misleading uncertainty quantification. This post reviews recent work on model misspecification in neural SBI, covering formal definitions, methods for detection and mitigation, and their underlying assumptions. It also discusses practical implications for SBI workflows and outlines open challenges for developing robust SBI methods that remain reliable in realistic, imperfectly specified applications.Poster SessionShowcase+
Format · ShowcasePoster SessionPresenters display work; attendees browse and ask questions. Some interaction, not structured work.
Evidence basisMediumRead from source
all eventLearning to Maximize Rewards via Reaching GoalsUnknownUnknown+
Format · BroadcastUnknownFormat not classified from the source; treated as a broadcast block by default.
Evidence basisLowRead from source
all eventDiscretisation invarianceUnknownUnknown+
Format · BroadcastUnknownFormat not classified from the source; treated as a broadcast block by default.
Evidence basisLowRead from source
all eventJustRL: Scaling a 1.5B LLM with a Simple RL RecipeUnknownUnknown+
Format · BroadcastUnknownFormat not classified from the source; treated as a broadcast block by default.
Evidence basisLowRead from source
all eventBudget Alignment: Making Models Reason in the User's LanguageUnknownUnknown+
Format · BroadcastUnknownFormat not classified from the source; treated as a broadcast block by default.
Evidence basisLowRead from source
all eventUnigramLM: An Attempt at Writing The Missing ManualUnknownUnknown+
Format · BroadcastUnknownFormat not classified from the source; treated as a broadcast block by default.
Evidence basisLowRead from source
all eventHeuristic-Based Ideation for Guiding LLMs Toward Structured CreativityUnknownUnknown+
Format · BroadcastUnknownFormat not classified from the source; treated as a broadcast block by default.
Evidence basisLowRead from source
all eventRevisiting the NetHack Learning EnvironmentUnknownUnknown+
Format · BroadcastUnknownFormat not classified from the source; treated as a broadcast block by default.
Evidence basisLowRead from source
all eventContent Promotion as a Strategic Game: How to Design Agentic Publishers for the Evolving Search Ecosystem in the GenAI Era?UnknownUnknown+
Format · BroadcastUnknownFormat not classified from the source; treated as a broadcast block by default.
Evidence basisLowRead from source
all eventUnderstanding and Fixing Bottlenecks in State Space Models: What Recency and Over-Smoothing Tell UsUnknownUnknown+
Format · BroadcastUnknownFormat not classified from the source; treated as a broadcast block by default.
Evidence basisLowRead from source
all eventSquare Peg, Round Hole: Plugging Non-Sequential Data into Sequential Language ModelsUnknownUnknown+
Format · BroadcastUnknownFormat not classified from the source; treated as a broadcast block by default.
Evidence basisLowRead from source
all eventReady For General Agents? Let's test it.UnknownUnknown+
Format · BroadcastUnknownFormat not classified from the source; treated as a broadcast block by default.
Evidence basisLowRead from source
all eventFlow Where You WantUnknownUnknown+
Format · BroadcastUnknownFormat not classified from the source; treated as a broadcast block by default.
Evidence basisLowRead from source
all eventDon't Look Up (Every Token): Escaping Quadratic Complexity via Geometric Patterns and AlgorithmsUnknownUnknown+
Format · BroadcastUnknownFormat not classified from the source; treated as a broadcast block by default.
Evidence basisLowRead from source
all eventMisalignments and RL Failure Modes in the Early Stage of SuperintelligenceUnknownUnknown+
Format · BroadcastUnknownFormat not classified from the source; treated as a broadcast block by default.
Evidence basisLowRead from source
all eventComputer Use Survey - A Visual Survey of Computer Use AgentsUnknownUnknown+
Format · BroadcastUnknownFormat not classified from the source; treated as a broadcast block by default.
Evidence basisLowRead from source
all eventProbabilistic Circuits for Uncertainty QuantificationUnknownUnknown+
Format · BroadcastUnknownFormat not classified from the source; treated as a broadcast block by default.
Evidence basisLowRead from source
all eventGenerative AI ArchaeologyUnknownUnknown+
Format · BroadcastUnknownFormat not classified from the source; treated as a broadcast block by default.
Evidence basisLowRead from source
all eventLoneliness as a Case Study for Social Reward MisalignmentUnknownUnknown+
Format · BroadcastUnknownFormat not classified from the source; treated as a broadcast block by default.
Evidence basisLowRead from source
all eventWhy AI Evaluations Need Error BarsUnknownUnknown+
Format · BroadcastUnknownFormat not classified from the source; treated as a broadcast block by default.
Evidence basisLowRead from source
all eventAI Fundamentals: Valuing AI Agents & Data AssetsUnknownUnknown+
Format · BroadcastUnknownFormat not classified from the source; treated as a broadcast block by default.
Evidence basisLowRead from source
The Full Reading
Why It Ranks This Way +
Calibrated from GES design 34/100 and verified 34/100 with no fourth-loop cap.
Reader Takeaway. For a reader, this is a comparison record more than a model to copy: it reads as a broadcast-heavy conference, with the strongest visible signal in learning transfer and problem specificity and the biggest open question around follow through and network design. The practical test is whether the published agenda connects the room to post-event continuation and evidence.
Strongest signals: Learning Transfer, Problem Specificity, and Evidence Maturity. Weakest signals: Follow Through, Network Design, and Participation Architecture.
How This Agenda Could Improve +
- Add named owners, dates, implementation checkpoints, and a visible post-event continuation path.
- Replace generic networking blocks with designed introductions, ask-offer exchanges, peer groups, or bridge-building rituals.
- Turn passive airtime into participant work: practice, sensemaking, decisions, critique, or artifact creation.
Fastest next move: Add named owners, dated next steps, and a visible continuation path before treating the event as outcome-ready.
Role-Specific Reading +
Sponsor lens
Look beyond exposure. Strong sponsor value would show qualified interaction, problem work, buyer learning, customer evidence, or follow-up. The practical sponsor move is to look for structured introductions, buyer-seller fit, and relationship persistence.
Designer lens
The agenda is useful as a pattern sample from iclr.cc. Redesign attention should go first to the lowest-scoring pillars; in practice, turn the thinnest agenda blocks into participant work.
Executive lens
Treat the visible agenda as an operating plan. The executive move is to require owners, dates, and evidence before treating the event as strategic. If owners, proof, and follow-through are not visible, the public record does not yet prove strategic movement.
Aggregator lens
Treat the source URL as evidence, not decoration. The data-product move is to label the source boundary clearly before ranking the record before ranking or syndicating the record.
What GES Means Here +
The Gathering Effectiveness Score is a strict 0-100 public-evidence reading of the agenda across eight pillars. It rewards visible participant work, follow-through, transfer, network design, and proof mechanisms more than polish, speaker fame, attendance, or satisfaction.
Visible mechanisms: Participant work, Feedback, Network design, Learning transfer, Personalization.
Evidence boundary: Scores reflect visible agenda/source evidence and should not be read as proof of causal event impact.
Limitations, Score Caps, and Review Flags +
Limitations
- No visible follow-up, progress monitoring, or longitudinal tracking.
- Passive stage formats dominate the visible agenda.
- No baseline measurement is visible.
Score caps
- No fourth-loop score cap applied.
Review flags
- No source-backed follow-up, validation, baseline, tracking, or impact evidence.
- No tracking, validation, feedback, or impact measurement found in the visible source text.
- Original category was unknown; publication category is inferred.
Is this proof the event worked? +
No. This is a strict public-evidence reading of the agenda. Proof would require baseline, comparison, follow-up, attribution, and impact evidence beyond the listing.
What should a reader inspect first? +
Start with the source URL, then compare the eight pillar scores against the agenda rows. The biggest opportunities usually sit in follow-through, evidence maturity, and participant work.
Why publish weak records? +
Weak records are part of the map. They show where public agendas still describe sessions and speakers more often than outcomes, commitments, transfer, or proof.
How should I use the rows? +
Read the agenda rows as the visible design trace: formats, purposes, and evidence labels show what the public source made inspectable, not everything that happened in the room. This is a source-grounded interpretation of the public agenda record, not a copy of the source, and not an endorsement of the event.
Where To Go Next
Compare this agenda against other Academic / Research / Science events scored on the same eight pillars.