Paradigm Shift

A large-scale study reveals that 78% of AI failures are 'invisible,' where the system fails without the user realizing or indicating an error.

Introduces an adversarial co-evolution framework where Code and Test LLMs optimize against each other to improve code generation.

This paper proposes a method to align and personalize LLMs directly from raw user interactions using self-distillation, bypassing the need for explicit human labels or RLHF.

Introduces the Budget-Sensitive Discovery Score (BSDS), a formally verified metric machine-checked in Lean 4 for evaluating AI-guided scientific candidate selection.

This paper establishes a systematic protocol for 'stitching' heterogeneous Vision Foundation Models (e.g., CLIP and DINOv2) to share early layers while retaining specialized capabilities.

Introduces Modal Logical Neural Networks (MLNNs) as a differentiable logic layer that bridges deep learning with symbolic Kripke semantics for regulated AI.

Demonstrates a robot that improves its own locomotion by identifying and physically 'self-destructing' redundant or inhibiting limbs during its lifetime.

Derives an exact, unbiased policy gradient for Reinforcement Learning on Diffusion LLMs, bypassing the need for sequence-level likelihood approximations.

Proposes modeling the world in the feature space of frozen geometry foundation models instead of pixels, achieving 5x faster depth forecasting.

A small-scale molecular reasoning model that outperforms ultra-large foundation models via structured chain-of-thought and RL.

ThinkStream introduces a 'Watch-Think-Speak' paradigm for video reasoning that allows models to incrementally update understanding and decide when to respond in real-time.

Connects DDIM reverse chains to fractal geometry, providing a mathematical explanation for why diffusion models switch from global context to local detail.

Proposes Causal Process Reward (CPR) to fix 'cherry-picking' in MLLM reasoning by coupling answer correctness with step-level logical alignment.

Reimagines 3D molecules as continuous vector fields rather than discrete graphs, decoupling structure learning from atom types.

This paper introduces a graph tokenization framework that allows standard Transformers like BERT to beat specialized Graph Neural Networks without any architectural changes.

Continual Representation Learning (CoRe) moves PEFT from weight-level updates to representation-space interventions, solving catastrophic forgetting in dynamic environments.

Theoretical analysis proves that Langevin dynamics is fundamentally non-robust to score function errors, justifying the shift to Diffusion Models.

HAPO resolves the advantage collapse problem in sparse-reward RL for reasoning models using a Thompson-sampled hindsight mechanism.

This paper introduces Finsler geometry to manifold learning, allowing for the capture of asymmetric data relationships like density hierarchies that Riemannian methods ignore.

Manifold-Optimal Guidance reformulates Classifier-Free Guidance (CFG) as a Riemannian control problem, eliminating the artifacts and saturation typical of high guidance scales.

Expert Threshold Routing (ET) replaces standard top-k token-choice with an independent thresholding mechanism, achieving 1.6x faster training convergence.

Introduces the Compression-Consistency Principle, arguing that LLMs prefer truth only when false alternatives are structurally harder to compress.

Enables agents to autonomously discover the group structure of their environments to learn disentangled representations without human priors.

Eliminates lookahead bias in financial backtesting through a series of yearly-partitioned pretrained LLMs.

Solves GNN over-squashing by using global effective resistance to identify and rewire structural bottlenecks.

Proposes a unified image tokenizer that reconciles the conflicting requirements of visual understanding and generation using a residual evolution process.

Introduces a feature-matching objective for LLM fine-tuning that targets sequence-level statistics without requiring reward models or ground-truth verifiers.

Demonstrates that the stochasticity in standard regularized model training (like cross-validation) can serve as a 'free' and effective exploration strategy for contextual bandits.