arXiv:2605.05226v1 Announce Type: new Abstract: The central challenge of reinforcement learning for reasoning lies not only in the sparsity of outcome-level supervision, but more fundamentally in how to transform feedback provided only at the end of a sequence into fine-grained…
arXiv:2508.14482v3 Announce Type: replace Abstract: The explainability of deep learning models remains a significant challenge, particularly in the medical domain where interpretable outputs are essential for clinical trust and transparency. Path attribution methods such as Integrated Gradients rely on a…
arXiv:2605.05227v1 Announce Type: new Abstract: Data curation is a critical yet under-explored area in large language model (LLM) training. Existing methods, such as data selection and mixing, operate in an offline paradigm, detaching themselves from training. This separation introduces engineering…
arXiv:2601.21351v2 Announce Type: replace Abstract: Attentio-FFN disaggregation (AFD) is an emerging architecture for LLM decoding that separates state-heavy, KV-cache-dominated Attention computation from stateless, compute-intensive FFN computation, connected by per-step communication. While AFD enables independent scaling of memory and compute resources,…
arXiv:2605.05228v1 Announce Type: new Abstract: Deep learning models are the most efficient models in many machine learning tasks. The main disadvantage when using them in IoT, mobile devices, independent autonomous or real-time systems is their complexity and memory size. Therefore,…
arXiv:2603.20991v2 Announce Type: replace Abstract: Compressing transformer weights makes large language models cheaper to deploy. But each layer’s compression introduces an error. These errors accumulate as the signal passes through later layers, and how they accumulate is not well understood.…
arXiv:2605.05278v1 Announce Type: new Abstract: Resource-efficient machine learning increasingly uses sparse Mixture-of-Experts (MoE) architectures, where the gate acts as both a learning component and a routing interface controlling computation, communication, and accuracy. Motivated by finite-rate interpretations of MoE gating, we…
arXiv:2605.01627v2 Announce Type: replace Abstract: Low-rank decomposition is a compelling approach for compressing large language models, but its effectiveness hinges on selecting which singular-vector bases to retain for a target task. Existing methods such as Basel adapt singular-value coefficients on…
arXiv:2412.10665v2 Announce Type: replace-cross Abstract: We introduce a foundation model for event classification in high-energy physics, built on a Graph Neural Network architecture and trained on 120 million simulated proton-proton collision events spanning 12 distinct physics processes. The model is…
arXiv:2602.07618v5 Announce Type: replace Abstract: We investigate the approximation capabilities of dense neural networks. While universal approximation theorems establish that sufficiently large architectures can approximate arbitrary continuous functions if there are no restrictions on the weight values, we show that…
