New Capability

Mamba-3 introduces MIMO formulations and complex-valued updates to solve the state-tracking failures of previous linear models.

Uses Sparse Autoencoders (SAEs) to mechanisticially repair 'moral indifference' in LLM latent representations.

A benchmark for unsolved math problems with automated verification, enabling the measurement of true mathematical discovery.

Enables Bayesian model selection and joint posterior inference over combinatorial spaces of up to billions of simulator model instantiations.

Enables training-free infinite video generation (hour-scale) by using evolving memory tokens to solve identity drift and motion stagnation.

Unlocks Maximum Entropy RL for high-dimensional humanoid control, matching or doubling the performance of dominant deterministic baselines.

A retrosynthesis model that explicitly learns strategic bond-disconnection reasoning via reinforcement learning with a round-trip accuracy reward.

A new system enables humanoid robots to play competitive tennis rallies with humans by learning from imperfect, fragmented motion data.

SciDesignBench provides a massive simulator-grounded environment for scientific inverse design, revealing that current LLMs struggle significantly with iterative refinement.

A self-supervised robotic system detects novel objects by training bespoke detectors on-the-fly from human video demonstrations, bypassing language-based prompts.

AIM enables post-training modulation of large models to change utility levels or focus features without any retraining or additional data.

First training-free method for debiasing reward models using Sparse Autoencoder (SAE) interventions.

A flow-based navigation policy that achieves zero-shot sim-to-real transfer across wheeled, quadrupedal, and humanoid platforms.

MotionAnymesh automatically transforms static 3D meshes into simulation-ready, articulated digital twins for robotics using vision-language models grounded in physical priors.

Multimodal OCR (MOCR) treats charts, diagrams, and tables as code-level targets (e.g., TikZ, SVG) rather than just cropping them as pixels.

Optimizes diffusion models via Direct Preference Optimization (DPO) to generate human motion that is inherently executable by real humanoid robots.

A new method for training axis-aligned decision trees using gradient descent and backpropagation, allowing trees to be integrated into end-to-end neural networks.

SoLA introduces the first reversible model editing framework that allows precise revocation of specific knowledge updates.

RewardHackingAgents establishes a benchmark for evaluating whether ML-engineering agents are actually solving tasks or just tampering with the evaluation code.

RoboClaw introduces 'Entangled Action Pairs' to allow robots to autonomously collect data by learning to reset their own environment.

Replaces unstructured LLM debates with 'Deliberative Collective Intelligence,' producing formal decision packets with minority reports and accountability trails.

Automates the entire robotic data generation loop, including a self-resetting mechanism that restores unstructured workspaces without human intervention.

Bridges the gap between parametric CAD and direct B-Rep synthesis using LLMs and primitive grounding.

Enables concurrent perception and reasoning for continuous video streams in Multimodal Large Language Models.

First framework for interpreting 4D molecular trajectories into natural language explanations.

Cross-domain sensor model that handles variable signal lengths and resolutions without retraining.

Enables multimodal agents to continually improve from experience and skills without any parameter updates through a dual-stream visual grounding framework.

A 3D vision-language pipeline that grounds medical diagnosis in longitudinal brain MRI via regional volumetric assessments to eliminate VLM hallucinations.

Integrates Neural ODEs with NeRFs to enable continuous-time scene dynamics that can extrapolate far beyond the original training sequence.

Integrates Chain-of-Thought reasoning directly into the Diffusion Transformer denoising process to solve complex spatial and logical tasks.

Enables VideoLLMs to perform complex logical reasoning simultaneously with video playback without incurring the latency of standard test-time scaling.

A unified streaming visual backbone that performs perception, 3D reconstruction, and robotic action simultaneously from a continuous video stream.