3/11/2026
Reasoning as Gradient: Scaling MLE Agents Beyond Tree Search
LLM-based agent (GOME) for machine learning engineering using gradient-based reasoning instead of tree search for efficient optimization.
3/11/2026
The Geometric Inductive Bias of Grokking: Bypassing Phase Transitions via Architectural Topology
Mechanistic interpretability study using architectural topology modifications to investigate grokking and delayed generalization in Transformers.
3/11/2026
Censored LLMs as a Natural Testbed for Secret Knowledge Elicitation
Study on eliciting truthful responses from censored LLMs and detecting deception using naturally-occurring dishonesty patterns.
3/11/2026
Omni-Masked Gradient Descent: Memory-Efficient Optimization via Mask Traversal with Improved Convergence
Memory-efficient optimization method for training large language models with mask traversal and convergence guarantees for nonconvex settings.
3/11/2026
Khatri-Rao Clustering for Data Summarization
Clustering algorithm for data summarization using Khatri-Rao product to reduce redundancies in centroid-based cluster prototypes.
3/11/2026
OptiRoulette Optimizer: A New Stochastic Meta-Optimizer for up to 5.3x Faster Convergence
Meta-optimizer that dynamically selects update rules during training with PyTorch compatibility, achieving 5.3x faster convergence.
3/11/2026
Property-driven Protein Inverse Folding With Multi-Objective Preference Alignment
Multi-objective protein sequence design balancing designability with competing properties like solubility and thermostability using preference alignment.
3/11/2026
Adversarial Latent-State Training for Robust Policies in Partially Observable Domains
Reinforcement learning approach for robust policies under latent distribution shift in partially observable domains using adversarial training.
3/11/2026
Latent Generative Models with Tunable Complexity for Compressed Sensing and other Inverse Problems
Generative models with tunable complexity for inverse problems, extending diffusion models and normalizing flows with adaptive dimensionality.
3/11/2026
FedPrism: Adaptive Personalized Federated Learning under Non-IID Data
Federated learning framework addressing non-IID data distribution through personalization strategies for heterogeneous client data.
3/11/2026
Impermanent: A Live Benchmark for Temporal Generalization in Time Series Forecasting
Benchmark for evaluating time-series forecasting foundation models with temporal generalization, addressing train-test contamination issues in static splits.
3/11/2026
Enhancing Computational Efficiency in Multiscale Systems Using Deep Learning of Coordinates and Flow Maps
Deep learning method for learning coordinates and flow maps to enhance computational efficiency in multiscale dynamical systems.
3/11/2026
DRUPI: Dataset Reduction Using Privileged Information
DRUPI: dataset reduction method using privileged information beyond input-label pairs for dataset condensation tasks.
3/11/2026
Learning responsibility allocations for multi-agent interactions: A differentiable optimization approach with control barrier functions
Differentiable optimization approach using control barrier functions to learn safe responsibility allocations in multi-agent autonomous systems.
3/11/2026
Adaptive and Stratified Subsampling for High-Dimensional Robust Estimation
Adaptive importance sampling and stratified subsampling estimators for robust high-dimensional sparse regression with heavy-tailed noise.
3/11/2026
Prognostics for Autonomous Deep-Space Habitat Health Management under Multiple Unknown Failure Modes
Prognostics system for autonomous deep-space habitat health monitoring and remaining useful life prediction under multiple failure modes.
3/11/2026
Morphological-Symmetry-Equivariant Heterogeneous Graph Neural Network for Robotic Dynamics Learning
MS-HGNN: morphological-symmetry-equivariant heterogeneous graph neural network for robotic dynamics learning with structural priors.
3/11/2026
CuriousBot: Interactive Mobile Exploration via Actionable 3D Relational Object Graph
CuriousBot: mobile robotic exploration system using actionable 3D relational object graphs for interactive environment exploration.
3/11/2026
A Distributional Treatment of Real2Sim2Real for Object-Centric Agent Adaptation in Vision-Driven Deformable Linear Object Manipulation
Real2Sim2Real framework using likelihood-free inference for visual robotic manipulation of deformable linear objects.
3/11/2026
Concept Drift Guided LayerNorm Tuning for Efficient Multimodal Metaphor Identification
LayerNorm tuning method using concept drift for efficient multimodal metaphor identification in internet memes.
3/11/2026
UltraEdit: Training-, Subject-, and Memory-Free Lifelong Editing in Language Models
UltraEdit: training-free model editing approach for lifelong knowledge updates in LLMs without retraining or subject-specific data.
3/11/2026
Cooperative Game-Theoretic Credit Assignment for Multi-Agent Policy Gradients via the Core
CORA: cooperative game-theoretic credit assignment method for multi-agent reinforcement learning using core concepts for advantage allocation.
3/11/2026
Regret-Optimal Q-Learning with Low Cost for Single-Agent and Federated Reinforcement Learning
Regret-optimal Q-learning algorithms minimizing sample collection and policy switching costs in single-agent and federated reinforcement learning.
3/11/2026
ConLID: Supervised Contrastive Learning for Low-Resource Language Identification
Supervised contrastive learning approach for low-resource language identification to improve multilingual LLM pretraining corpus curation.
3/11/2026
Convergence Rate for the Last Iterate of Stochastic Gradient Descent Schemes
Research on convergence rates for stochastic gradient descent and heavy ball methods under convex and non-convex objectives using discrete Gronwall's inequality.
3/11/2026
Latent Policy Steering with Embodiment-Agnostic Pretrained World Models
Latent policy steering approach using embodiment-agnostic pretrained world models to leverage cross-embodiment datasets for robot learning.
3/11/2026
Robot Control Stack: A Lean Ecosystem for Robot Learning at Scale
Robot Control Stack ecosystem for scalable vision-language-action model training and deployment, replacing traditional robotics frameworks.
3/11/2026
Compose Your Policies! Improving Diffusion-based or Flow-based Robot Policies via Test-time Distribution-level Composition
Test-time composition method for enhancing diffusion-based robot control policies without additional training data or model fine-tuning.
3/11/2026
Latent Speech-Text Transformer
Latent Speech-Text Transformer improving compute efficiency of speech-text models by reducing token sequence length through latent representations.
3/11/2026
AlphaApollo: A System for Deep Agentic Reasoning
AlphaApollo agentic reasoning system addressing reasoning capacity and verification bottlenecks through multi-turn reasoning and trustworthy tool integration.
3/11/2026
Exploring Single Domain Generalization of LiDAR-based Semantic Segmentation under Imperfect Labels
Domain generalization approach for LiDAR semantic segmentation handling imperfect labels and sensor noise in autonomous driving scenarios.
3/11/2026
RECODE: Reasoning Through Code Generation for Visual Question Answering
RECODE agentic framework using code generation and derendering for verifiable visual reasoning on structured visuals with multimodal LLMs.
3/11/2026
RL-100: Performant Robotic Manipulation with Real-World Reinforcement Learning
RL-100 real-world robotic manipulation framework combining diffusion visuomotor policies with imitation and reinforcement learning via clipped PPO.
3/11/2026
Personalized Collaborative Learning with Affinity-Based Variance Reduction
Personalized collaborative learning framework with affinity-based variance reduction for heterogeneous multi-agent systems.
3/11/2026
From Spatial to Actions: Grounding Vision-Language-Action Model in Spatial Foundation Priors
FALCON vision-language-action model incorporating 3D spatial foundation priors to improve reasoning and generalization in embodied AI tasks.
3/11/2026
An Interpretable Operator-Learning Model for Electric Field Profile Reconstruction in Discharges Based on the EFISH Method
Interpretable operator-learning ML model for reconstructing electric field distributions from EFISH signal profiles in plasma physics.
3/11/2026
Fairness-Aware Fine-Tuning of Vision-Language Models for Medical Glaucoma Diagnosis
Fairness-aware LoRA fine-tuning of vision-language models for medical imaging with differentiable MaxAccGap loss for demographic parity optimization.
3/11/2026
Enhancing Retrieval-Augmented Generation with Entity Linking for Educational Platforms
ELERAG system enhancing retrieval-augmented generation with entity linking to improve factual accuracy in specialized domains like education.
3/11/2026
ADHint: Adaptive Hints with Difficulty Priors for Reinforcement Learning
ADHint method integrating difficulty-aware hints into reinforcement learning post-training to improve sample efficiency and reasoning generalization.
3/11/2026
Provable Acceleration of Distributed Optimization with Local Updates
Theoretical analysis of distributed optimization with multiple local updates between communication rounds, proving acceleration guarantees.
3/11/2026
An AI-powered Bayesian Generative Modeling Approach for Arbitrary Conditional Inference
Bayesian generative modeling framework for flexible conditional inference on arbitrary partitions of observed variables without fixed conditioning structure constraints.
3/11/2026
Latent Equivariant Operators for Robust Object Recognition: Promises and Challenges
Equivariant neural networks for robust object recognition under symmetric transformations and unusual viewing conditions.
3/11/2026
FinTexTS: Financial Text-Paired Time-Series Dataset via Semantic-Based and Multi-Level Pairing
FinTexTS dataset pairs financial time-series with semantic text data for multi-modal financial forecasting and analysis.
3/11/2026
Reasoning Theater: Disentangling Model Beliefs from Chain-of-Thought
Analysis of performative chain-of-thought in reasoning models, showing models generate tokens without revealing internal beliefs via activation probing.
3/11/2026
PolyBlocks: A Compiler Infrastructure for AI Chips and Programming Frameworks
PolyBlocks: MLIR-based modular compiler infrastructure for AI programming frameworks and chips using affine analysis and analytical cost models.
3/11/2026
VLN-Cache: Enabling Token Caching for VLN Models with Visual/Semantic Dynamics Awareness
VLN-Cache improves vision-language model inference efficiency for Vision-and-Language Navigation via semantic-aware token caching.
3/11/2026
Scalable Training of Mixture-of-Experts Models with Megatron Core
Megatron Core system optimizations for scaling Mixture-of-Experts model training across memory, communication, and computation constraints.
3/11/2026
Covenant-72B: Pre-Training a 72B LLM with Trustless Peers Over-the-Internet
Covenant-72B: 72B parameter LLM trained via globally distributed, trustless peer-to-peer training over the internet without whitelisting.
3/11/2026
A prospective clinical feasibility study of a conversational diagnostic AI in an ambulatory primary care clinic
Clinical feasibility study of AMIE, an LLM-based conversational AI for patient diagnostic history in real-world primary care workflows.
3/11/2026
PostTrainBench: Can LLM Agents Automate LLM Post-Training?
PostTrainBench benchmarks LLM agents' ability to automate post-training of language models, extending AI agents to AI research automation.