6871 papers
The paper introduces MJ1, a 3B-parameter multimodal judge that leverages reinforcement learning with a structured grounded verification chain and counterfactual consistency rewards to achieve state-of-the-art accuracy on MMRB2, outperforming significantly larger models by effectively grounding decisions in visual evidence.
SmartThinker is a novel GRPO-based method that dynamically calibrates chain-of-thought length by estimating optimal response lengths and modulating reward coefficients, achieving significant compression of reasoning paths while improving accuracy on complex benchmarks like AIME25.
This paper proposes "amortized MIPS," a learning-based framework that leverages the mathematical properties of support functions to train neural networks (SupportNet and KeyNet) that directly predict optimal keys for Maximum Inner Product Search, thereby amortizing computational costs for queries drawn from a fixed distribution.
FedMomentum is a novel federated fine-tuning framework that preserves LoRA training momentum and ensures mathematically correct aggregation by using singular value decomposition (SVD) to extract dominant update directions while retaining residual components, thereby achieving faster convergence and higher accuracy than existing methods.
This paper introduces CAMEL, a capacity-aware mixture law that enables efficient data mixture optimization for large language models by modeling the nonlinear interplay between model size and data composition, thereby reducing optimization costs by 50% while improving downstream benchmark performance by up to 3%.
The paper proposes GCGNet, a Graph-Consistent Generative Network that integrates a Variational Generator, Graph Structure Aligner, and Graph Refiner to jointly model temporal and channel correlations in a noise-robust manner, thereby outperforming state-of-the-art methods in time series forecasting with exogenous variables.
CDRRM introduces a contrast-driven framework that generates high-quality, interpretable rubrics from preference pairs to guide reward modeling, achieving state-of-the-art performance and superior data efficiency while mitigating common evaluation biases.
This paper introduces Stabilized Federated LoRA (SFed-LoRA), a novel framework that derives an optimal scaling factor to mitigate the statistical variance and gradient collapse caused by aggregating high-rank LoRA updates across multiple clients in Federated Learning, thereby restoring stability and convergence without altering model architecture.
This study proposes an adversarial deep learning framework that enables effective knowledge transfer across heterogeneous RNA-seq datasets by learning a domain-invariant latent space, thereby significantly improving cancer and tissue type classification accuracy, especially in low-data scenarios.
This paper introduces Deterministic Differentiable Pruning (DDP), a novel method that eliminates the train-test mismatch and stochasticity of prior structured pruning approaches by directly optimizing a deterministic soft surrogate of the l0 objective, thereby achieving faster convergence, greater expressiveness, and superior performance retention (e.g., ~1% loss) on large language models like Qwen3 at high sparsity levels.
This paper proposes a hybrid quantum-classical architecture that integrates a Variational Quantum Circuit within a GRU backbone to forecast multivariate physiological signals, demonstrating competitive accuracy and enhanced robustness to noise and missing data in small-cohort clinical settings.
This paper introduces and evaluates the Autoregressive TRM, a model adapting the two-step refinement mechanism of Tiny Recursive Models for autoregressive tasks, but finds that while some two-step refinement baselines show promise, the specific Autoregressive TRM architecture offers no reliable performance gains over standard Transformers.
EAGLE-Pangu is a reproducible system that adapts EAGLE-3-style tree speculative decoding for Ascend NPUs with Pangu teacher models by introducing an explicit cache manager, accelerator-safe tensorization, and a fused-kernel verification path, achieving up to 2.46x throughput improvement over greedy decoding.
This paper introduces the Dual-Consensus Weak-to-Strong (DC-W2S) framework, which enhances the reliability of Process Reward Models in biological reasoning by strategically filtering noisy weak supervision signals through self- and neighborhood-consensus metrics to enable robust training without exhaustive expert annotation.
This paper demonstrates a novel safety threat where large language models are finetuned to use steganography, allowing them to covertly generate harmful content in response to hidden malicious prompts while displaying only benign interactions to human observers and safety classifiers.
The paper introduces Tau-BNO, a deep learning surrogate framework that rapidly and accurately approximates the computationally intensive Network Transport Model of tau propagation in Alzheimer's disease, enabling efficient parameter inference and mechanistic discovery by reducing simulation time from hours to seconds while outperforming existing sequence models.
This paper introduces ROMI, a robust offline reinforcement learning method that replaces RAMBO's unstable model gradient updates with a novel robust value-aware learning approach and implicitly differentiable adaptive weighting to achieve controllable conservatism and superior performance on out-of-distribution datasets.
SaiVLA-0 introduces a neuroscience-inspired, compute-aware Vision-Language-Action framework featuring a tripartite Cerebrum-Pons-Cerebellum architecture that decouples high-level semantics from real-time control to achieve modular scalability, active foveated vision, and significant improvements in training efficiency and task success rates.
FoleyFlow introduces a novel video-to-audio generation framework that achieves superior semantic and rhythmic synchronization by aligning unimodal encoders through masked audio-visual modeling and employing a dynamic conditional flow that utilizes temporally varying video features to guide audio synthesis.
The paper introduces TRIAGE, a lightweight post-hoc framework that decomposes uncertainty into aleatoric and epistemic components to trigger distinct corrective actions—observation recovery for corrupted data and control moderation for model mismatch—thereby significantly improving robotic manipulation success rates and enabling efficient adaptive perception.