New Capability

Demonstrates a complete AI-assisted mathematical research loop where a mathematician wrote zero lines of formal code to verify complex physics equilibria.

Integrates LLM agents with the industry-standard Rosetta software to automate physics-based protein design for non-canonical amino acids.

Enables the prediction of an adapter's task, performance, and attributes directly from its LoRA weights without any inference or data access.

Introduces ARISE, a hierarchical reinforcement learning framework that allows LLMs to evolve and reuse a tiered library of reasoning skills rather than treating every math problem in isolation.

Proposes the Vision-Sound-Language-Action (VSLA) paradigm, enabling robots to respond to real-time environmental acoustics during task execution.

Successfully trains a 0.9B parameter pure Spiking Neural Network (SNN) from scratch for language modeling, achieving performance without Transformer distillation.

Localizes reinforcement learning updates for code generation by using execution traces to identify the exact point of semantic failure.

Uses an asymmetric Draft-Verify-Recover pipeline to enable high-quality personalized AI assistants without compromising user privacy.

A self-supervised RLVR method that escapes the 'spurious majority' trap by using a temporary unlearning process for exploration.

Omnilingual MT scales machine translation to over 1,600 languages, an 8x increase in coverage over previous state-of-the-art systems.

This paper demonstrates precise behavioral steering of agentic traits in a 35B parameter MoE model using Sparse Autoencoder (SAE) decoded probe vectors.

Introduces a method to give frozen LLMs persistent memory in their continuous latent space, bypassing the need for text-level RAG or retraining.

Capability-Guided Compression uses Sparse Autoencoders (SAEs) to prevent 'capability loss' during model pruning and quantization.

Detects and mitigates Vision-Language Model hallucinations at inference time by analyzing visual attention entropy rather than text outputs.

Introduces a way to train Reward Models that generate 'transferable rubrics'—explicit scoring criteria that improve performance across different tasks and models.

OmniSONAR scales cross-lingual sentence embeddings to over 1,500 languages across text, speech, code, and math in a single semantic space.

Fine-tuning language models on journal publication records allows them to match or exceed human experts in judging 'scientific taste'—the ability to identify which research ideas are worth pursuing.

This method non-rigidly aligns inconsistent video diffusion frames into globally-consistent 3D pointclouds to enable high-quality environment reconstruction.

pADAM is a unified generative framework that learns shared priors across heterogeneous multi-physics families (e.g., scalar diffusion to Navier-Stokes).

SOMA provides a unified, differentiable layer that bridges incompatible human body models like SMPL and SMPL-X in a single closed-form pass.

LEAFE allows LLM agents to internalize feedback as actionable experience, enabling them to backtrack and recover from failures autonomously.

Prism prevents 'diversity collapse' in self-evolving reasoning systems by using semantic partitioning to guide the generation of new problems.

Safety fine-tuning causes representational collapse in the residual stream, leading to 'false refusals' of benign queries.

By fine-tuning on categorical refusal tokens, researchers can extract steerable directions to control fine-grained refusal behavior during inference.

Latent Entropy-Aware Decoding (LEAD) mitigates hallucinations by switching between discrete token and continuous probability-weighted embeddings based on real-time uncertainty.

Introduces event-gated sampling to eliminate interaction hallucinations in video generation, such as objects drifting after placement.

Uses generative world models to synthesize photorealistic, counterfactual failure data for training robot recovery behaviors.

Introduces StatePlane, a model-agnostic memory architecture that enables long-horizon AI reasoning without expanding the context window or KV cache.

KoopmanFlow uses a Koopman-inspired structural bias to decouple global steady-state motions from high-frequency local corrections in robotic control policies.

GradMem replaces the massive KV-cache with a compact memory state updated via test-time gradient descent.

Proposes URDF-Anything+, an autoregressive framework that generates fully executable articulated 3D models from raw visual observations.

Introduces the first system capable of imaging high-speed, non-rigid objects through strong atmospheric turbulence at 16,000 pixels per second.

Enables online, incremental 3D Gaussian Splatting for thousands of frames by replacing global reprocessing with a causal, streaming update framework.

Introduces a decentralized, multi-agent framework for scientific discovery that uses an 'ArtifactReactor' for plannerless coordination and full computational lineage.

Introduces 'Visual Chronometer' to estimate physical frame rates directly from visual dynamics, addressing the 'chronometric hallucinations' common in generative video models.

Segment Anything Reasoner (StAR) successfully introduces parallel test-time scaling to visual segmentation tasks, eliciting latent reasoning capabilities from base models.

V-JEPA 2.1 unlocks dense, spatially structured features in video self-supervised learning, yielding massive gains in robotic manipulation and navigation.

One-Policy-Fits-All (OPFA) learns a single manipulation policy across 11 different embodiments, including grippers and dexterous hands, using geometry-aware action latents.

Interp3R is the first method to estimate depth and camera poses at arbitrary time instants by interpolating pointmaps using asynchronous event data.

MorFiC achieves zero-shot locomotion transfer across quadrupeds of different sizes and masses with up to 5x speed gains over standard baselines.

Discovers interpretable 'atoms' of model behavior by decomposing training gradients, enabling unsupervised discovery and steering of complex behaviors like refusal or arithmetic.

Achieves pose-free 3D Gaussian Splatting using only event streams, enabling reconstruction in extreme lighting and high-speed motion scenarios.

A training-free operator for streaming 3D reconstruction reduces geometric drift using Grassmannian manifolds.

DynaAvatar achieves zero-shot 3D human reconstruction from a single image with motion-dependent cloth dynamics.

Euler Characteristic Surfaces achieve 98% accuracy on time-series classification with O(n) complexity, crushing previous topological methods that only hit 62%.

ForceVLA2 introduces explicit force awareness and hybrid control to Vision-Language-Action models, enabling stable contact-rich manipulation.

SCAN enables reliable sequential knowledge editing in LLMs for up to 3,000 edits without the catastrophic forgetting or model collapse seen in current methods.

This physics-informed VLM framework improves physics-grounded anomaly detection AUROC from 66.9% to 96.7%.

FuXiWeather2 is a unified end-to-end neural framework for weather assimilation and forecasting that outperforms global operational systems.

Incorporating PDE residuals into fine-tuning allows pre-trained physics foundation models to adapt to new tasks without requiring ground-truth solutions.