arXiv:2602. 21321v2 Announce Type: replace Abstract: Analog in-memory computing (AIMC) performs computation directly within resistive crossbar arrays, offering an energy-efficient platform to scale large vision and language models.
arXiv:2609.36584v1 Announce Type: new Abstract: Analog in-memory computing (AIMC) offers an alternative for model training by executing matrix operations directly where weights are stored. However, s...
arXiv:2607. 28418v1 Announce Type: cross Abstract: Pruning is a promising approach for improving the efficiency of LLMs.
TinyUDE introduces a solver‑free training framework for Universal Differential Equations (UDEs) that fits a hybrid vector field directly to the first and second time‑derivatives of observed system states using Lie‑Taylor jet matching. The method estimates these derivatives online via Savitzky‑Golay filtering, enabling fully analytic gradients without automatic differentiation. Experiments on damped and chaotic pendulum systems show that TinyUDE matches or surpasses conventional solver‑based baselines while using less than 108 kB of static memory and achieving real‑time on‑device training on an ESP32 microcontroller.
arXiv:2607. 18306v1 Announce Type: cross Abstract: Sharpness-Aware Minimization (SAM) improves generalization by minimizing the worst-case loss in a local parameter neighborhood.
arXiv:2606. 15004v1 Announce Type: cross Abstract: Deploying neural networks on low-power microcontrollers (MCUs) requires selecting model architectures under tight memory, latency, and energy constraints.
arXiv:2606. 17249v1 Announce Type: cross Abstract: The dominant trajectory of modern machine learning has been to scale up: larger models, larger accelerators, larger memory budgets.
arXiv:2607. 11585v1 Announce Type: cross Abstract: Low-Gain Avalanche Diodes (LGADs) and AC-coupled Low-Gain Avalanche Diodes (AC-LGADs) are promising technologies for precision timing and four-dimensional tracking.
arXiv:2608. 02700v1 Announce Type: cross Abstract: Analog compute-in-memory (CIM) enables energy-efficient neural network inference, but device variation and read noise can severely degrade low-bit quantized models.
arXiv:2606. 26795v1 Announce Type: cross Abstract: Video Diffusion Models (VDMs) is constrained by immense computational costs.
The paper introduces dattri-LLM, a library designed to make training data attribution (TDA) practical for large language models. It achieves efficiency by using compact gradient representations and a cost‑based routing system, while maintaining compatibility by capturing per‑example gradients from existing training loops without modifications, even in distributed settings. The library also offers extensibility through reusable gradient operations and callbacks, supporting various attribution methods and applications such as online data selection, and demonstrates significant performance gains and scalability up to 110B‑parameter models.
arXiv:2607. 03349v1 Announce Type: cross Abstract: The accuracy of consumer-grade inertial navigation is bottlenecked by the stochastic noise of Micro-Electro-Mechanical Systems (MEMS).
arXiv:2607. 21999v1 Announce Type: new Abstract: Long-tailed learning couples two sources of poor generalization: head classes dominate training exposure, while under-represented classes often converge to sharper regions of the loss landscape.