Explore documented scientific discoveries, original research and new scientific developments. Search the library by topic, category or date.
arXiv:2610.40998v1 Announce Type: new Abstract: Steering vectors (SVs) are residual-stream directions that shift the behavior of LLMs when added to hidden states at inference time. While prior work has mainly used SVs as post-hoc controls, we ask whether the same directions can also guide parameter updates during post-training.
arXiv MLRead article →arXiv:2610.40984v1 Announce Type: new Abstract: A nonlinear dynamical system can be characterized by the autonomous evolution of its state in time and its dependence on system parameters. As a parameter varies, the dynamics can change quantitatively and, at bifurcation points, qualitatively.
arXiv MLRead article →arXiv:2610.40956v1 Announce Type: new Abstract: We introduce the Monte Carlo Random Walk Kernel (MCRWK), an estimator whose per-pair cost is independent of graph size and derived directly from the random walk kernel (RWK) expansion. By normalizing the kernel coefficients, MCRWK samples a shared walk length and estimates its contribution using random walks on the input graphs.
arXiv MLRead article →arXiv:2610.40954v1 Announce Type: new Abstract: Recent progress in deep learning has substantially advanced automated grading of ophthalmologic pathologies, supported by large annotated datasets and improved representation learning. However, a significant gap remains between state-of-the-art performance and clinical scalability.
arXiv MLRead article →arXiv:2610.40945v1 Announce Type: new Abstract: On-policy reinforcement learning for large language models assumes that training responses are sampled from the current policy. In practice, an inference engine samples the responses and a separate trainer computes the gradients, and the \textbf{training-inference mismatch} between the two makes the training data off-policy.
arXiv MLRead article →arXiv:2610.40943v1 Announce Type: new Abstract: The bottleneck of attention mechanisms, which are the fundamental building block of modern language models, is their quadratic complexity in terms of context size. There has thus been an increasing interest in the compression of the context, also called KV-cache compression.
arXiv MLRead article →arXiv:2610.40936v1 Announce Type: new Abstract: Recommendation systems that optimize for the average user can fail to capture the preferences of diverse demographic groups. We consider a dynamic recommendation system, where arriving users can belong to different demographic groups. Importantly, these groups can intersect, and a user may belong to several groups.
arXiv MLRead article →arXiv:2610.40934v1 Announce Type: new Abstract: In this work, we provide a formal framework and mathematical definition for the subset of foundation models focused on enabling the property of "any-task inference" (ATI) (similar to zero-shot inference), rather than those designed for pre-training fine-tuning.
arXiv MLRead article →arXiv:2610.40914v1 Announce Type: new Abstract: Training acceleration methods for diffusion transformers, such as REPA and dispersive regularization, are commonly explained by additional supervision, spatial-structure priors, or improved feature geometry.
arXiv MLRead article →arXiv:2610.40901v1 Announce Type: new Abstract: As publications and products on time series forecasting (TSF) proliferate, researchers face a growing burden of analyzing existing work. First, not all publications release code. For instance, only about 20% of arXiv TSF papers have a code link in the surveyed period. Second, released code also requires manual maintenance.
arXiv MLRead article →arXiv:2610.40871v1 Announce Type: new Abstract: High-dimensional tensors are ubiquitous in real-world applications but lead to massive storage overhead. Functional tensor decomposition addresses this by employing neural networks to map element coordinates to tensor cores, significantly reducing the parameters for data compression.
arXiv MLRead article →arXiv:2610.40861v1 Announce Type: new Abstract: We study first-order regret guarantees for adversarial bandits with a fixed, known directed feedback graph. For strongly observable graphs, [ACBDK15] establishes a minimax regret of $\widetilde{\Theta}(\sqrt{\alpha T})$, where $\alpha$ denotes the independence number of the graph and $T$ denotes the horizon.
arXiv MLRead article →arXiv:2610.40858v1 Announce Type: new Abstract: Training a large language model is compute-bound, and current recipes divide the work between optimizers by hand: orthogonalized momentum for attention and MLP matrices, Adam-style updates for the rest. This division rests on human priors about which layer type deserves which geometry, and it fixes a single global update budget that every parameter block shares.
arXiv MLRead article →arXiv:2610.40857v1 Announce Type: new Abstract: Scientific foundation models (SciFMs) have achieved remarkable performance on a wide range of PDEs, including the ability to efficiently adapt to unseen data distributions with additional finetuning.
arXiv MLRead article →arXiv:2610.40757v1 Announce Type: new Abstract: Humans and animals have a unique capacity to generalize and learn in novel circumstances, including "0-shot" inference (ZI) and continual learning (CL) -- two powerful cognitive abilities. Despite this importance, efforts to model ZI and CL are seriously limited in several ways.
arXiv MLRead article →arXiv:2610.40740v1 Announce Type: new Abstract: Continual adaptation requires deciding not only how to update a model, but whether to commit an update without exceeding a declared budget for changes to protected behavior. We introduce Certified Anchor-Based Adaptation (CABA), a candidate-agnostic post-update gate that reserves historical observations from candidate optimization, compares old and candidate policies…
arXiv MLRead article →arXiv:2610.40736v1 Announce Type: new Abstract: High dimensionality degrades both the efficiency and the effectiveness of learning algorithms, and in many applications no class labels are available to guide dimensionality reduction. This paper proposes LIDFS, an unsupervised feature selection method based on Local Intrinsic Dimensionality (LID).
arXiv MLRead article →arXiv:2610.40732v1 Announce Type: new Abstract: Industrial Internet of Things (IIoT) edge pipelines routinely lose contiguous blocks of multi-sensor data to network dropouts and hardware faults. The standard remedy is a disjointed "impute-then-detect" pipeline whose imputer is trained only for Mean Squared Error (MSE).
arXiv MLRead article →arXiv:2610.40728v1 Announce Type: new Abstract: Medium-voltage feeders have few real-time measurements, so distribution system state estimation (DSSE) rests on pseudo-measurements of the secondary substations, a prior in Bayesian terms.
arXiv MLRead article →arXiv:2610.40700v1 Announce Type: new Abstract: To rigorously establish the value of a novel deep learning method, it must be compared against appropriate baselines. As deep learning has become increasingly common in the physical and life sciences, there have been numerous cases in which apparent progress has turned out to be an artifact of omitted or missing baselines.
arXiv MLRead article →arXiv:2607.16408v2 Announce Type: replace-cross Abstract: Idle network service cores are treated as wasted compute. This assumption motivates increasingly sophisticated mechanisms that reclaim idle cores at microsecond timescales. We argue that this view no longer matches modern server hardware. On contemporary multicore processors, active cores compete for a shared package-level power and thermal budget.
arXiv Computer Science ARRead article →arXiv:2610.11748v2 Announce Type: replace Abstract: U-Net inference for brain-tumor segmentation requires billions of multiply-accumulate operations, motivating hardware that can reduce computation dynamically rather than relying only on fixed precision or static model compression.
arXiv Computer Science ARRead article →arXiv:2604.26103v3 Announce Type: replace Abstract: All current LLM serving systems place the GPU at the center, from production-level attention-FFN disaggregation to NVIDIA's Rubin GPU-LPU heterogeneous platform. Even academic PIM/PNM proposals still treat the GPU as the central hub for cross-device communication.
arXiv Computer Science ARRead article →arXiv:2610.41919v1 Announce Type: cross Abstract: We evaluate NVIDIA B200 and B300 (Blackwell Ultra) GPUs on two independent eight-GPU clusters, spanning computer-vision inference and training, LLM inference with Llama-3-8B and 70B~\cite{llama3}, and six production HPC codes.
arXiv Computer Science ARRead article →