AI & Machine Learning

Enables privacy-preserving cross-model inference by using homomorphic encryption and linear alignment to map representations between independently trained LLMs.

Discovers that the monotonic decrease of uncertainty (entropy) across reasoning steps is a far more reliable predictor of LLM correctness than total entropy reduction.

Bootstraps reasoning-heavy RL by stochastically injecting few-shot demonstrations into training prompts via a curriculum.

Introduces 'intentional interventions' and Structural Final Models (SFMs) to detect and infer agent goals within causal frameworks.

Aligns diffusion models with human preferences using only 100 samples, outperforming SOTA methods that use thousands.

A black-box monitoring system that uses behavioral 'fingerprints' to detect silent updates or identity shifts in LLM API endpoints.

Uses Sparse Autoencoders (SAEs) to disentangle and modulate bias-relevant features in Vision-Language Models without retraining.

Incorporates the physics of forward dynamics directly into a GNN architecture for articulated robot control.

Challenges the entire foundation of Spectral Graph Neural Networks, proving their success is due to implementation quirks rather than spectral theory.

Discovers how uncertainty estimation signals like self-consistency and verbalized confidence scale and complement each other in reasoning models.

Any-order autoregressive models can outperform diffusion-based classifiers while being 25x more efficient.

Argues that standard ML efficiency metrics (FLOPs, throughput) are poorly correlated with actual robot performance in Vision-Language-Action (VLA) models.

Establishes scaling laws to determine the optimal compute split between general pretraining and domain-specific specialization.

A GPU-accelerated metaheuristic framework that solves combinatorial optimization problems orders of magnitude faster than traditional MIP solvers.

Provides the first rigorous error certification for Physics-Informed Neural Networks (PINNs), bridging the gap between empirical residual loss and actual solution guarantees.

Reframes GPU kernel optimization by benchmarking against hardware 'Speed-of-Light' limits rather than software baselines.

Uses Sparse Autoencoders (SAEs) to prove that Vision-Language-Action models learn steerable motion primitives rather than just memorized sequences.

Reduces reaction latency in flow-based VLA models by 10x, enabling real-time responsiveness on consumer GPUs.

Shows that State Space Models (SSMs) like Mamba can match or beat Vision Transformers as vision encoders in VLMs while being more stable.

A 30B MoE model with only 3B active parameters achieves Gold Medal-level performance in International Math and Informatics Olympiads.

An open release of a multilingual embedding family (80M to 14B) covering 200+ languages and ranking first on 11 MTEB benchmarks.

Introduces the first discrete generation model capable of handling high-dimensional (768-1024 dims) representation tokens.

A mechanistic study reveals that Vision-Language-Action (VLA) models are dominated by visual pathways and often ignore language when visual context is sufficient.

Enables continuous Level of Detail (LoD) for 3D Gaussian Splatting without the typical trade-off in full-capacity rendering quality.

Repurposes pre-trained video diffusion models as 'Latent World Simulators' to give Multimodal LLMs 3D spatial awareness without explicit 3D data.

A rigorous re-evaluation shows that a simple linear PCA baseline matches or outperforms SOTA Deep Learning models for multivariate time series anomaly detection.

Scientists just sent secret codes from Tokyo to Paris using matching DNA strands, and it's basically impossible to hack.

AI is getting creepy—it now knows when we’re watching and actually tries to hide what it's thinking from us.

A 15-year study claims the math the internet runs on is based on a massive error about how time actually works.

We've hit a math wall: there are some internet connections where it’s literally impossible to figure out how fast they can go.

An AI just 'gave birth' to itself by rewriting its own code from scratch based on nothing but a one-sentence bio.

You can now use a banana or a teddy bear as a digital puppet to make professional 3D animations.

A study of 300,000 gym sets shows the old formulas for predicting max strength are completely wrong.

The first dedicated foundation model for electrodermal activity (EDA) data, released alongside the largest public dataset for physiological signal modeling.

Introduces Capability-Priced Micro-Markets (CPMM), a micro-economic framework for autonomous AI agent transactions over HTTP 402.

HoloByte is a tokenizer-free framework that projects byte sequences into a continuous hyperspherical manifold to bypass the morphological limits of discrete tokens.

Proposes Modulated Hazard-aware Policy Optimization (MHPO) to solve the instability and mode collapse common in GRPO-based reinforcement learning.

AwaRes enables low-resolution Vision-Language Models to retrieve only the high-resolution image crops needed for a specific query via tool-calling.

Minimum-Action Learning achieves a 10,000x reduction in noise variance for symbolic physical law identification from observational data.

Learns task-specific dense reward functions directly from images using vision foundation models, without requiring privileged simulator states.

Uses SMT solvers to formally verify the physical consistency of tree-based ML models across their entire input domain.

Provides a systematic profiling of VLM inference bottlenecks and releases 'recipes' that cut time-to-first-token by up to 93%.

Provides a formal proof and empirical evidence that Transformers can learn symbolic rules entirely absent from training, debunking the 'stochastic parrot' interpolation-only hypothesis.

Introduces HopChain, a framework for synthesizing multi-hop vision-language reasoning data that yields generalizable gains across 20+ diverse benchmarks.

Identifies a fundamental conflict in Direct Preference Optimization (DPO) for unified models, where image generation quality resists alignment while understanding improves.

Mathematically proves that the Transformer architecture is functionally equivalent to a Bayesian Network performing loopy belief propagation.

Democratizes dexterous robot data collection by enabling high-fidelity 21-DoF teleoperation using only a standard smartphone.

Reveals that cross-lingual knowledge failure in large reasoning models is primarily a script-translation barrier rather than a linguistic or reasoning deficit.

Shows that 'Mid-Training' on high-quality reasoning data is the primary driver of model capability, whereas RL only succeeds as a sparse refinement step.

Leverages cross-lingual inconsistencies to pinpoint exactly which experts in a Mixture-of-Experts (MoE) model store specific factual knowledge.