NeurIPS 2025 Schedule

Filter Events

SAT 29 NOV

1 p.m.

Registration Desk

(ends 4:00 PM)

SUN 30 NOV

6 a.m.

Registration Desk

(ends 4:00 PM)

7 a.m.

Break:

Coffee Break

(ends 7:30 AM)

Affinity Workshop:

Women in ML, LatinX in AI, Queer in AI Affinity workshop

(ends 11:00 AM)

10:30 a.m.

Break:

Coffee Break

(ends 11:30 AM)

11 a.m.

Workshop:

NeurIPS 2025 Workshop on Socially Responsible and Trustworthy Foundation Models

(ends 6:00 PM)

Affinity Poster Session:

Women in ML, LatinX in AI, Queer in AI Affinity Poster Session

(ends 1:00 PM)

Workshop:

Vision Language Models: Challenges of Real World Deployment

(ends 6:00 PM)

Workshop:

NeurIPS 2025 Workshop on Embodied and Safe-Assured Robotic Systems

(ends 6:00 PM)

Workshop:

NeurIPS2025 Workshop Research Development of AI in Mexico: Main Applications

(ends 6:00 PM)

2:30 p.m.

Break:

Coffee Break

(ends 3:30 PM)

MON 1 DEC

5 a.m.

Breakfast:

Breakfast

(ends 8:00 AM)

5:30 a.m.

Registration Desk

(ends 3:00 PM)

6 a.m.

Streaming Lounge [6:00-4:00]

(ends 4:00 PM)

6:30 a.m.

Workshop:

First Workshop on LLM Persona Modeling

(ends 1:30 PM)

6:50 a.m.

Workshop:

7th International Workshop on Large Scale Holistic Video Understanding: Toward Video Foundation Models

(ends 10:40 AM)

7 a.m.

Workshop:

NORA: The First Workshop on Knowledge Graphs & Agentic Systems Interplay

(ends 1:45 PM)

7:30 a.m.

Workshop:

Centering Low-Resource Languages and Cultures in the Age of Large Language Models

(ends 3:00 PM)

8 a.m.

Break:

Coffee Break

(ends 8:45 AM)

10 a.m.

Break:

Lunch

(ends 11:00 AM)

1 p.m.

Break:

Coffee Break

(ends 1:45 PM)

4 p.m.

TUE 2 DEC

5 a.m.

Breakfast:

Breakfast

(ends 8:00 AM)

9 a.m.

Registration Desk

(ends 3:00 PM)

Break:

Coffee Break

(ends 10:00 AM)

9:30 a.m.

Tutorial:

From Tuning to Guarantees: Statistically Valid Hyperparameter Selection

(ends 12:00 PM)

Tutorial:

Efficient Transformers: State of the art in pruning, sparse attention, and transformer funneling

(ends 12:00 PM)

Tutorial:

Geospatial Foundation Models: Overview, Application and Benchmarking

(ends 12:00 PM)

Streaming Lounge [9:30-4:00]

(ends 4:00 PM)

noon

Break:

Lunch

(ends 1:00 PM)

1:30 p.m.

Tutorial:

Positional Encoding: Past, Present, and Future

(ends 4:00 PM)

Tutorial:

Science of Trustworthy Generative Foundation Models

(ends 4:00 PM)

Tutorial:

How to Build Agents to Generate Kernels for Faster LLMs (and Other Models!)

(ends 4:00 PM)

4 p.m.

Reception:

Opening Reception

(ends 6:00 PM)

WED 3 DEC

5 a.m.

Breakfast:

Breakfast

(ends 8:00 AM)

8 a.m.

Registration Desk

(ends 4:00 PM)

Break:

Coffee Break

(ends 8:45 AM)

8:30 a.m.

Streaming Lounge [8:30-4:00]

(ends 4:00 PM)

Mexico City Invited Talk:

The Oak Architecture: A Vision of SuperIntelligence from Experience

Rich Sutton

(ends 9:30 AM)

9:30 a.m.

Break:

Lunch

(ends 10:30 AM)

10 a.m.

Oral 1A Language Model 1 [10:00-11:00]

Orals 10:00-11:00

[10:00] Understanding and Mitigating Numerical Sources of Nondeterminism in LLM Inference

[10:20] Artificial Hivemind: The Open-Ended Homogeneity of Language Models (and Beyond)

[10:40] SAVVY: Spatial Awareness via Audio-Visual LLMs through Seeing and Hearing

(ends 11:00 AM)

Oral 1B Generation/Simulation 1 [10:00-11:00]

Orals 10:00-11:00

[10:00] Representation Entanglement for Generation: Training Diffusion Transformers Is Much Easier Than You Think

[10:20] On the Closed-Form of Flow Matching: Generalization Does Not Arise from Target Stochasticity

[10:40] Why Diffusion Models Don’t Memorize: The Role of Implicit Dynamical Regularization in Training

(ends 11:00 AM)

Oral 1C Theory 1 [10:00-11:00]

Orals 10:00-11:00

[10:00] Optimal Mistake Bounds for Transductive Online Learning

[10:20] High-Dimensional Calibration from Swap Regret

[10:40] Does Stochastic Gradient really succeed for bandits?

(ends 11:00 AM)

Oral 1D Multimodal 1 [10:00-11:00]

Orals 10:00-11:00

[10:00] Dynam3D: Dynamic Layered 3D Tokens Empower VLM for Vision-and-Language Navigation

[10:20] Perception Encoder: The best visual embeddings are not at the output of the network

[10:40] Interactive Cross-modal Learning for Text-3D Scene Retrieval

(ends 11:00 AM)

11 a.m.

Mexico City Poster Session 1 [11:00-2:00]

Posters 11:00-2:00

Majority of the Bests: Improving Best-of-N via Bootstrapping

NegoCollab: A Common Representation Negotiation Approach for Heterogeneous Collaborative Perception

Token Perturbation Guidance for Diffusion Models

Reconstructing Heterogeneous Biomolecules via Hierarchical Gaussian Mixtures and Part Discovery

Beyond Scalar Rewards: An Axiomatic Framework for Lexicographic MDPs

Tight Lower Bounds and Improved Convergence in Performative Prediction

Abstract Counterfactuals for Language Model Agents

SPARKE: Scalable Prompt-Aware Diversity and Novelty Guidance in Diffusion Models via RKE Score

$\epsilon$-Seg: Sparsely Supervised Semantic Segmentation of Microscopy Data

$\texttt{STRCMP}$: Integrating Graph Structural Priors with Language Models for Combinatorial Optimization

Adaptive Sigmoid Clipping for Balancing the Direction–Magnitude Mismatch Trade-off in Differentially Private Learning

A Learning-Augmented Approach to Online Allocation Problems

Attack by Yourself: Effective and Unnoticeable Multi-Category Graph Backdoor Attacks with Subgraph Triggers Pool

AugGen: Synthetic Augmentation using Diffusion Models Can Improve Recognition

BADiff: Bandwidth Adaptive Diffusion Model

BNMusic: Blending Environmental Noises into Personalized Music

Can Class-Priors Help Single-Positive Multi-Label Learning?

Capturing Individual Human Preferences with Reward Features

Care-PD: A Multi-Site Anonymized Clinical Dataset for Parkinson’s Disease Gait Assessment

ChemX: A Collection of Chemistry Datasets for Benchmarking Automated Information Extraction

CHPO: Constrained Hybrid-action Policy Optimization for Reinforcement Learning

Constant Bit-size Transformers Are Turing Complete

DERD-Net: Learning Depth from Event-based Ray Densities

Enhancing Sample Selection Against Label Noise by Cutting Mislabeled Easy Examples

Enhancing Temporal Understanding in Video-LLMs through Stacked Temporal Attention in Vision Encoders

Exploring Semantic-constrained Adversarial Example with Instruction Uncertainty Reduction

Feature Distillation is the Better Choice for Model-Heterogeneous Federated Learning

Global Minimizers of $\ell^p$-Regularized Objectives Yield the Sparsest ReLU Neural Networks

Gradient Variance Reveals Failure Modes in Flow-Based Generative Models

GRAVER: Generative Graph Vocabularies for Robust Graph Foundation Models Fine-tuning

Learning Simple Interpolants for Linear Integer Arithmetic

Manipulating Feature Visualizations with Gradient Slingshots

Mixtures of Subspaces for Bandwidth Efficient Context Parallel Training

MoniTor: Exploiting Large Language Models with Instruction for Online Video Anomaly Detection

Non-convex entropic mean-field optimization via Best Response flow

One Sample is Enough to Make Conformal Prediction Robust

On the VC dimension of deep group convolutional neural networks

Physics-informed Reduced Order Modeling of Time-dependent PDEs via Differentiable Solvers

PointMAC: Meta-Learned Adaptation for Robust Test-Time Point Cloud Completion

Predictable Scale (Part II) --- Farseer: A Refined Scaling Law in LLMs

Relaxing partition admissibility in Cluster-DAGs: a causal calculus with arbitrary variable clustering

scSplit: Bringing Severity Cognizance to Image Decomposition in Fluorescence Microscopy

Spatial Understanding from Videos: Structured Prompts Meet Simulation Data

Strassen Attention, Split VC Dimension and Compositionality in Transformers

The Promise of RL for Autoregressive Image Editing

TreeSplat: Mergeable Tree for Deformable Gaussian Splatting

TRUST: Test-Time Refinement using Uncertainty-Guided SSM Traverses

Uncertainty-Aware Multi-Objective Reinforcement Learning-Guided Diffusion Models for 3D De Novo Molecular Design

Understanding Prompt Tuning and In-Context Learning via Meta-Learning

(ends 2:00 PM)

2 p.m.

Break:

Coffee Break

(ends 2:45 PM)

Test Of Time:

Test of Time Award

(ends 2:30 PM)

2:30 p.m.

Mexico City Invited Talk:

Are We Having the Wrong Nightmares About AI?

Zeynep Tufekci

(ends 3:30 PM)

3:30 p.m.

Oral 2A Multimodal 2 [3:30-4:30]

Orals 3:30-4:30

[3:30] Envisioning Beyond the Pixels: Benchmarking Reasoning-Informed Visual Editing

[3:50] CoralVQA: A Large-Scale Visual Question Answering Dataset for Coral Reef Image Understanding

[4:10] OpenHOI: Open-World Hand-Object Interaction Synthesis with Multimodal Large Language Model

(ends 4:30 PM)

Oral 2B Deep Learning 1 [3:30-4:30]

Orals 3:30-4:30

[3:30] A multiscale analysis of mean-field transformers in the moderate interaction regime

[3:50] The emergence of sparse attention: impact of data distribution and benefits of repetition

[4:10] From Condensation to Rank Collapse: A Two-Stage Analysis of Transformer Training Dynamics

(ends 4:30 PM)

Oral 2C Reinforcement/State-space 1 [3:30-4:30]

Orals 3:30-4:30

[3:30] PRIMT: Preference-based Reinforcement Learning with Multimodal Feedback and Trajectory Synthesis from Foundation Models

[3:50] Adaptive Surrogate Gradients for Sequential Reinforcement Learning in Spiking Neural Networks

[4:10] SAGE: A Unified Framework for Generalizable Object State Recognition with State-Action Graph Embedding

(ends 4:30 PM)

Oral 2E Theory 2 [3:30-4:30]

Orals 3:30-4:30

[3:30] Agnostic Active Learning Is Always Better Than Passive Learning

[3:50] Dynamical Decoupling of Generalization and Overfitting in Large Two-Layer Networks

[4:10] Tighter CMI-Based Generalization Bounds via Stochastic Projection and Quantization

(ends 4:30 PM)

4:30 p.m.

Startup Pitch:

NeurIPS 2025 Mexico City: Startup Pitch Competition

(ends 5:30 PM)

Mexico City Poster Session 2 [4:30-7:30]

Posters 4:30-7:30

Transcending Cost-Quality Tradeoff in Agent Serving via Session-Awareness

3DID: Direct 3D Inverse Design for Aerodynamics with Physics-Aware Optimization

Implicit Modeling for Transferability Estimation of Vision Foundation Models

FedIGL: Federated Invariant Graph Learning for Non-IID Graphs

Pessimistic Data Integration for Policy Evaluation

Flow-Based Policy for Online Reinforcement Learning

Test-Time Adaptive Object Detection with Foundation Model

Sampling-Efficient Test-Time Scaling: Self-Estimating the Best-of-N Sampling in Early Decoding

SYMPHONY: Synergistic Multi-agent Planning with Heterogeneous Language Model Assembly

PolypSense3D: A Multi-Source Benchmark Dataset for Depth-Aware Polyp Size Measurement in Endoscopy

No Object Is an Island: Enhancing 3D Semantic Segmentation Generalization with Diffusion Models

Adaptive Batch-Wise Sample Scheduling for Direct Preference Optimization

Addressing Mark Imbalance in Integration-free Marked Temporal Point Processes

An Evidence-Based Post-Hoc Adjustment Framework for Anomaly Detection Under Data Contamination

AutoPartGen: Autoregressive 3D Part Generation and Discovery

Detecting Data Deviations in Electronic Health Records

Diffusion-Guided Graph Data Augmentation

Don’t Give Up on Democratizing AI for the Wrong Reasons

Dr. RAW: Towards General High-Level Vision from RAW with Efficient Task Conditioning

Efficient Knowledge Transfer in Federated Recommendation for Joint Venture Ecosystem

Empirical Study on Robustness and Resilience in Cooperative Multi-Agent Reinforcement Learning

Enhancing GUI Agent with Uncertainty-Aware Self-Trained Evaluator

FrameShield: Adversarially Robust Video Anomaly Detection

Hierarchical Self-Attention: Generalizing Neural Attention Mechanics to Multi-Scale Problems

Identifying Macro Causal Effects in C-DMGs over DMGs

Learning to Plan Like the Human Brain via Visuospatial Perception and Semantic-Episodic Synergistic Decision-Making

Localized Data Shapley: Accelerating Valuation for Nearest Neighbor Algorithms

NAUTILUS: A Large Multimodal Model for Underwater Scene Understanding

OpenGU: A Comprehensive Benchmark for Graph Unlearning

PCA++: How Uniformity Induces Robustness to Background Noise in Contrastive Learning

PlanarGS: High-Fidelity Indoor 3D Gaussian Splatting Guided by Vision-Language Planar Priors

PolyMath: Evaluating Mathematical Reasoning in Multilingual Contexts

Statistical Parity with Exponential Weights

SubTrack++ : Gradient Subspace Tracking for Scalable LLM Training

Tighter CMI-Based Generalization Bounds via Stochastic Projection and Quantization

Towards Unified Multimodal Interleaved Generation via Group Relative Policy Optimization

Visual Structures Help Visual Reasoning: Addressing the Binding Problem in LVLMs

VPO: Reasoning Preferences Optimization Based on $\mathcal{V}$-Usable Information

World Models as Reference Trajectories for Rapid Motor Adaptation

(ends 7:30 PM)

THU 4 DEC

5 a.m.

Breakfast:

Breakfast

(ends 8:00 AM)

8 a.m.

Registration Desk

(ends 4:00 PM)

Break:

Coffee Break

(ends 8:45 AM)

8:30 a.m.

Mexico City Invited Talk:

The Art of (Artificial) Reasoning

Yejin Choi

(ends 9:30 AM)

Streaming Lounge [8:30-4:00]

(ends 4:00 PM)

9:30 a.m.

Break:

Lunch

(ends 10:30 AM)

10 a.m.

Oral 3A Reinforcement/State-space 2 [10:00-11:00]

Orals 10:00-11:00

[10:00] State Entropy Regularization for Robust Reinforcement Learning

[10:20] A Clean Slate for Offline Reinforcement Learning

[10:40] Breaking the Performance Ceiling in Reinforcement Learning requires Inference Strategies

(ends 11:00 AM)

Oral 3B Deep Learning 2 [10:00-11:00]

Orals 10:00-11:00

[10:00] Auto-Compressing Networks

[10:20] Dynamical Low-Rank Compression of Neural Networks with Robustness under Adversarial Attacks

[10:40] ImageNet-trained CNNs are not biased towards texture: Revisiting feature reliance through controlled suppression

(ends 11:00 AM)

Oral 3C Position Paper Orals [10:00-11:00]

Orals 10:00-11:00

[10:00] Position: If Innovation in AI systematically Violates Fundamental Rights, Is It Innovation at All?

[10:20] More effort is needed to protect pedestrian privacy in the era of AI

[10:40] Real-Time Hyper-Personalized Generative AI Should Be Regulated to Prevent the Rise of "Digital Heroin"

(ends 11:00 AM)

Oral 3D Application 2 [10:00-11:00]

Orals 10:00-11:00

[10:00] ControlFusion: A Controllable Image Fusion Network with Language-Vision Degradation Prompts

[10:20] Pan-LUT: Efficient Pan-sharpening via Learnable Look-Up Tables

[10:40] FuXi-Ocean: A Global Ocean Forecasting System with Sub-Daily Resolution

(ends 11:00 AM)

11 a.m.

Mexico City Poster Session 3 [11:00-2:00]

Posters 11:00-2:00

Dynamic Diffusion Schrödinger Bridge in Astrophysical Observational Inversions

Near-Optimal Experiment Design in Linear non-Gaussian Cyclic Models

Online Multi-Class Selection with Group Fairness Guarantee

Test-Time Adaptation of Vision-Language Models for Open-Vocabulary Semantic Segmentation

When Kernels Multiply, Clusters Unify: Fusing Embeddings with the Kronecker Product

StelLA: Subspace Learning in Low-rank Adaptation using Stiefel Manifold

Orthogonal Contrastive Learning for Multi-Representation fMRI Analysis

Hawaii: Hierarchical Visual Knowledge Transfer for Efficient Vision-Language Models

A machine learning approach that beats Rubik's cubes

A unified framework for establishing the universal approximation of transformer-type architectures

Beyond Pairwise Connections: Extracting High-Order Functional Brain Network Structures under Global Constraints

Computational Hardness of Reinforcement Learning with Partial $q^{\pi}$-Realizability

Doubly Robust Alignment for Large Language Models

ECO: Evolving Core Knowledge for Efficient Transfer

Enhancing Privacy in Multimodal Federated Learning with Information Theory

Fair Matroid Selection

Fast Rate Bounds for Multi-Task and Meta-Learning with Different Sample Sizes

Global Prompt Refinement with Non-Interfering Attention Masking for One-Shot Federated Learning

HoloLLM: Multisensory Foundation Model for Language-Grounded Human Sensing and Reasoning

IMPACT: Irregular Multi-Patch Adversarial Composition Based on Two‑Phase Optimization

Improving the Straight-Through Estimator with Zeroth-Order Information

Information Theoretic Learning for Diffusion Models with Warm Start

InstructHOI: Context-Aware Instruction for Multi-Modal Reasoning in Human-Object Interaction Detection

Limitations of Normalization in Attention

Multi-Objective Hyperparameter Selection via Hypothesis Testing on Reliability Graphs

MUVR: A Multi-Modal Untrimmed Video Retrieval Benchmark with Multi-Level Visual Correspondence

One-Step is Enough: Sparse Autoencoders for Text-to-Image Diffusion Models

Proximalized Preference Optimization for Diverse Feedback Types: A Decomposed Perspective on DPO

PT-MoE: An Efficient Finetuning Framework for Integrating Mixture-of-Experts into Prompt Tuning

Revealing Multimodal Causality with Large Language Models

Shapley-Based Data Valuation for Weighted $k$-Nearest Neighbors

Soft-consensual Federated Learning for Data Heterogeneity via Multiple Paths

SSRB: Direct Natural Language Querying to Massive Heterogeneous Semi-Structured Data

Statistical inference for Linear Stochastic Approximation with Markovian Noise

The Atlas of In-Context Learning: How Attention Heads Shape In-Context Retrieval Augmentation

Toward a Unified Geometry Understanding : Riemannian Diffusion Framework for Graph Generation and Prediction

Train with Perturbation, Infer after Merging: A Two-Stage Framework for Continual Learning

Trust Region Reward Optimization and Proximal Inverse Reward Optimization Algorithm

Unifying and Enhancing Graph Transformers via a Hierarchical Mask Framework

V2V: Scaling Event-Based Vision through Efficient Video-to-Voxel Simulation

(ends 2:00 PM)

2 p.m.

Break:

Coffee Break

(ends 2:45 PM)

2:30 p.m.

Mexico City Invited Talk:

On the Science of “Alien Intelligences”: Evaluating Cognitive Capabilities in Babies, Animals, and AI

Melanie Mitchell

(ends 3:30 PM)

3:30 p.m.

Oral 4A Language Model 2 [3:30-4:30]

Orals 3:30-4:30

[3:30] A is for Absorption: Studying Feature Splitting and Absorption in Sparse Autoencoders

[3:50] Gated Attention for Large Language Models: Non-linearity, Sparsity, and Attention-Sink-Free

[4:10] Superposition Yields Robust Neural Scaling

(ends 4:30 PM)

Oral 4B Generation/Simulation 2 [3:30-4:30]

Orals 3:30-4:30

[3:30] Exploring Diffusion Transformer Designs via Grafting

[3:50] Deep Compositional Phase Diffusion for Long Motion Sequence Generation

[4:10] Mean Flows for One-step Generative Modeling

(ends 4:30 PM)

Oral 4C Position Paper Panels [3:30-4:30]

Responsible AI Research & Unlearning: From Consent to Compliance to Critique

Strengthening the AI Research Ecosystem: Integrity, Critique, and Consensus

(ends 4:30 PM)

Oral 4D Optimization [3:30-4:30]

Orals 3:30-4:30

[3:30] In Search of Adam’s Secret Sauce

[3:50] Analog In-memory Training on General Non-ideal Resistive Elements: The Impact of Response Functions

[4:10] Generalized Gradient Norm Clipping & Non-Euclidean $(L_0,L_1)$-Smoothness

(ends 4:30 PM)

4:30 p.m.

Mexico City Poster Session 4 [4:30-7:30]

Posters 4:30-7:30

Breakthrough Sensor-Limited Single View: Towards Implicit Temporal Dynamics for Time Series Domain Adaptation

Explainably Safe Reinforcement Learning

GeoLLaVA-8K: Scaling Remote-Sensing Multimodal Large Language Models to 8K Resolution

How Different from the Past? Spatio-Temporal Time Series Forecasting with Self-Supervised Deviation Learning

Learning Grouped Lattice Vector Quantizers for Low-Bit LLM Compression

Leveraging Importance Sampling to Detach Alignment Modules from Large Language Models

Leveraging robust optimization for llm alignment under distribution shifts

LLM at Network Edge: A Layer-wise Efficient Federated Fine-tuning Approach

Local Curvature Descent: Squeezing More Curvature out of Standard and Polyak Gradient Descent

MSTAR: Box-free Multi-query Scene Text Retrieval with Attention Recycling

PhysDrive: A Multimodal Remote Physiological Measurement Dataset for In-vehicle Driver Monitoring

Pin the Tail on the Model: Blindfolded Repair of User-Flagged Failures in Text-to-Image Services

RankMatch: A Novel Approach to Semi-Supervised Label Distribution Learning Leveraging Rank Correlation between Labels

Reasoning Is Not a Race: When Stopping Early Beats Going Deeper

Robust Federated Finetuning of LLMs via Alternating Optimization of LoRA

Robust Graph Condensation via Classification Complexity Mitigation

Rope to Nope and Back Again: A New Hybrid Attention Strategy

S-Crescendo: A Nested Transformer Weaving Framework for Scalable Nonlinear System in S-Domain Representation

SPICED: A Synaptic Homeostasis-Inspired Framework for Unsupervised Continual EEG Decoding

Spurious-Aware Prototype Refinement for Reliable Out-of-Distribution Detection

Trained Mamba Emulates Online Gradient Descent in In-Context Linear Regression

Unsupervised Federated Graph Learning

(ends 7:30 PM)

FRI 5 DEC

5 a.m.

Breakfast:

Breakfast

(ends 8:00 AM)

8 a.m.

Registration Desk

(ends 12:00 PM)

Break:

Coffee Break

(ends 8:45 AM)

8:30 a.m.

Mexico City Invited Talk:

From Benchmarks to Problems - A Perspective on Problem Finding in AI

Kyunghyun Cho

(ends 9:30 AM)

Streaming Lounge [8:30-4:00]

(ends 4:00 PM)

9:30 a.m.

Break:

Lunch

(ends 10:30 AM)

10 a.m.

Oral 5A Language Model 3 [10:00-11:00]

Orals 10:00-11:00

[10:00] EvoLM: In Search of Lost Language Model Training Dynamics

[10:20] Large Language Diffusion Models

[10:40] Does Reinforcement Learning Really Incentivize Reasoning Capacity in LLMs Beyond the Base Model?

(ends 11:00 AM)

Oral 5B Multimodel 3 [10:00-11:00]

Orals 10:00-11:00

[10:00] Boosting Knowledge Utilization in Multimodal Large Language Models via Adaptive Logits Fusion and Attention Reallocation

[10:20] HyperET: Efficient Training in Hyperbolic Space for Multi-modal Large Language Models

[10:40] Rethinking Multimodal Learning from the Perspective of Mitigating Classification Ability Disproportion

(ends 11:00 AM)

11 a.m.

Mexico City Poster Session 5 [11:00-2:00]

Posters 11:00-7:30

AngleRoCL: Angle-Robust Concept Learning for Physically View-Invariant Adversarial Patches

User-Instructed Disparity-aware Defocus Control

HypoBootstrap: A Bootstrapping Framework for Inductive Reasoning

When Less Language is More: Language-Reasoning Disentanglement Makes LLMs Better Multilingual Reasoners

Teaching Language Models to Evolve with Users: Dynamic Profile Modeling for Personalized Alignment

SHAP values via sparse Fourier representation

Reward-Aware Proto-Representations in Reinforcement Learning

SSIMBaD: Sigma Scaling with SSIM-Guided Balanced Diffusion for AnimeFace Colorization

Connectome-Based Modelling Reveals Orientation Maps in the Drosophila Optic Lobe

Aligning Evaluation with Clinical Priorities: Calibration, Label Shift, and Error Costs

3D Gaussian Flats: Hybrid 2D/3D Photometric Scene Reconstruction

A Closer Look at Graph Transformers: Cross-Aggregation and Beyond

AdaLRS: Loss-Guided Adaptive Learning Rate Search for Efficient Foundation Model Pretraining

AI Progress Should Be Measured by Capability-Per-Resource, Not Scale Alone: A Framework for Gradient-Guided Resource Allocation in LLMs

Bridging Time and Linguistics: LLMs as Time Series Analyzer through Symbolization and Segmentation

COALA: Numerically Stable and Efficient Framework for Context-Aware Low-Rank Approximation

Compressed and Smooth Latent Space for Text Diffusion Modeling

CymbaDiff: Structured Spatial Diffusion for Sketch-based 3D Semantic Urban Scene Generation

DrivingRecon: Large 4D Gaussian Reconstruction Model For Autonomous Driving

GLNCD: Graph-Level Novel Category Discovery

Light-Weight Diffusion Multiplier and Uncertainty Quantification for Fourier Neural Operators

Memory-Augmented Potential Field Theory: A Framework for Adaptive Control in Non-Convex Domains

Mesh Interpolation Graph Network for Dynamic and Spatially Irregular Global Weather Forecasting

More Than Generation: Unifying Generation and Depth Estimation via Text-to-Image Diffusion Models

Mysteries of the Deep: Role of Intermediate Representations in Out of Distribution Detection

Navigating the MIL Trade-Off: Flexible Pooling for Whole Slide Image Classification

Nearly-Linear Time Private Hypothesis Selection with the Optimal Approximation Factor

Rectifying Soft-Label Entangled Bias in Long-Tailed Dataset Distillation

Reduction-based Pseudo-label Generation for Instance-dependent Partial Label Learning

Resource-Constrained Federated Continual Learning: What Does Matter?

Sample-Efficient Tabular Self-Play for Offline Robust Reinforcement Learning

Second-order Optimization under Heavy-Tailed Noise: Hessian Clipping and Sample Complexity Limits

THD-BAR: Topology Hierarchical Derived Brain Autoregressive Modeling for EEG Generic Representations

The Hawthorne Effect in Reasoning Models: Evaluating and Steering Test Awareness

The Quest for Universal Master Key Filters in DS-CNNs

Tight Bounds on the Distortion of Randomized and Deterministic Distributed Voting

Towards foundational LiDAR world models with efficient latent flow matching

Tru-POMDP: Task Planning Under Uncertainty via Tree of Hypotheses and Open-Ended POMDPs

UniDomain: Pretraining a Unified PDDL Domain from Real-World Demonstrations for Generalizable Robot Task Planning

YEAST: Yet Another Sequential Test

(ends 2:00 PM)

2 p.m.

Award:

Sejnowski-Hinton Award

(ends 2:30 PM)

Break:

Coffee Break

(ends 2:45 PM)

2:30 p.m.

Mexico City Invited Talk:

Demystifying depth: Principles of learning in deep neural networks

Andrew Saxe

(ends 3:30 PM)

3:30 p.m.

Oral 6A Reinforcement/State-space 3 [3:30-4:30]

Orals 3:30-4:30

[3:30] A Snapshot of Influence: A Local Data Attribution Framework for Online Reinforcement Learning

[3:50] 1000 Layer Networks for Self-Supervised RL: Scaling Depth Can Enable New Goal-Reaching Capabilities

[4:10] Learning long range dependencies through time reversal symmetry breaking

(ends 4:30 PM)

Oral 6B Multimodal 4 [3:30-4:30]

Orals 3:30-4:30

[3:30] KVzip: Query-Agnostic KV Cache Compression with Context Reconstruction

[3:50] MokA: Multimodal Low-Rank Adaptation for MLLMs

[4:10] ElasticMM: Efficient Multimodal LLMs Serving with Elastic Multimodal Parallelism

(ends 4:30 PM)

4:30 p.m.

Mexico City Poster Session 6 [4:30-7:30]

Posters 4:30-

LoRA-EnVar: Parameter-Efficient Hybrid Ensemble Variational Assimilation for Weather Forecasting

Afterburner: Reinforcement Learning Facilitates Self-Improving Code Efficiency Optimization

F-Adapter: Frequency-Adaptive Parameter-Efficient Fine-Tuning in Scientific Machine Learning

SPARTAN: A Sparse Transformer World Model Attending to What Matters

Neural Attention Search

A Dynamic Learning Strategy for Dempster-Shafer Theory with Applications in Classification and Enhancement

AffordBot: 3D Fine-grained Embodied Reasoning via Multimodal Large Language Models

Aligning What Matters: Masked Latent Adaptation for Text-to-Audio-Video Generation

Cognitive Predictive Processing: A Human-inspired Framework for Adaptive Exploration in Open-World Reinforcement Learning

Diversity Is All You Need for Contrastive Learning: Spectral Bounds on Gradient Magnitudes

EAReranker: Efficient Embedding Adequacy Assessment for Retrieval Augmented Generation

GraphKeeper: Graph Domain-Incremental Learning via Knowledge Disentanglement and Preservation

Integral Imprecise Probability Metrics

Linguini: A benchmark for language-agnostic linguistic reasoning

More Than Just Functional: LLM-as-a-Critique for Efficient Code Generation

OCRBench v2: An Improved Benchmark for Evaluating Large Multimodal Models on Visual Text Localization and Reasoning

Prompt Tuning Transformers for Data Memorization

RankSEG-RMA: An Efficient Segmentation Algorithm via Reciprocal Moment Approximation

Semi-supervised Graph Anomaly Detection via Robust Homophily Learning

SEMPO: Lightweight Foundation Models for Time Series Forecasting

Storyboard-guided Alignment for Fine-grained Video Action Recognition

The Quotient Bayesian Learning Rule

Turning the Tables: Enabling Backward Transfer via Causal-Aware LoRA in Continual Learning

UFO-RL: Uncertainty-Focused Optimization for Efficient Reinforcement Learning Data Selection

UniGTE: Unified Graph–Text Encoding for Zero-Shot Generalization across Graph Tasks and Domains

(ends 7:30 PM)