arXiv:2602.01381v1 Announce Type: cross Abstract: Inference-time scaling has recently emerged as a powerful paradigm for improving the reasoning capability of large language models. Among various approaches, Sequential Monte Carlo (SMC) has become a particularly important framework, enabling iterative generation, evaluation, rejection, and resampling of intermediate reasoning trajectories. A central component in this process is the reward model, which evaluates partial solutions and guides the allocation of computation during inference. However, in practice, true reward models are never available. […]

Ver mais

arXiv:2601.19908v1 Announce Type: new Abstract: The proliferation of large language models (LLMs) is accelerating the integration of multimodal assistants into edge devices, where inference is executed under stringent latency and energy constraints, often exacerbated by intermittent connectivity. These challenges become particularly acute in the context of multimodal LLMs (MLLMs), as high-dimensional visual inputs are transformed into extensive token sequences, thereby inflating the key-value (KV) cache and imposing substantial data movement overheads to the LLM backbone. To address these […]

Ver mais

Multivariable optimization is an essential mathematical exercise in daily engineering design and troubleshooting. To this end, simplex multivariable optimization method is a powerful optimization approach that has served many engineering disciplines well over the years. One such simplex algorithm is the Nelder Mead algorithm. But, the reflection step of the Nelder Mead algorithm may increase the number of iterations needed to arrive at the optimal point, as it reflects the starting point to the opposite side of the […]

Ver mais

arXiv:2509.24914v2 Announce Type: replace Abstract: Trained attention layers exhibit striking and reproducible spectral structure of the weights, including low-rank collapse, bulk deformation, and isolated spectral outliers, yet the origin of these phenomena and their implications for generalization remain poorly understood. We study empirical risk minimization in a single-head tied-attention layer trained on synthetic high-dimensional sequence tasks generated from the attention-indexed model. Using tools from random matrix theory, spin-glass theory, and approximate message passing, we obtain an exact high-dimensional […]

Ver mais

Introduction Disclaimer: This article is a cross post from Pytorch Medium Blog Post. One of the most exciting advancements, that has pushed the frontier of the Artificial Intelligence (AI) in recent years, is Deep Reinforcement Learning (DRL). DRL belongs to the family of machine learning algorithms. It assumes that intelligent machines can learn from their actions similar to the way humans learn from experience. Over the recent years we could witness some impressive real-world applications of DRL. The […]

Ver mais

arXiv:2601.20868v1 Announce Type: new Abstract: Large Language Models (LLMs) have advanced the field of Combinatorial Optimization through automated heuristic generation. Instead of relying on manual design, this LLM-Driven Heuristic Design (LHD) process leverages LLMs to iteratively generate and refine solvers to achieve high performance. However, existing LHD frameworks face two critical limitations: (1) Endpoint-only evaluation, which ranks solvers solely by final quality, ignoring the convergence process and runtime efficiency; (2) High adaptation costs, where distribution shifts necessitate re-adaptation […]

Ver mais

TL;DR I tested 112 Product Hunt startups with 2,240 queries across ChatGPT and Perplexity. The results challenge conventional wisdom about “Generative Engine Optimization” (GEO): The Discovery Gap: ChatGPT recognizes 99% of products directly but recommends only 3% organically (30:1 ratio) GEO doesn’t work (yet): Zero correlation between GEO optimization and ChatGPT discovery Traditional SEO wins: Backlinks (r=+0.32) and Reddit presence (r=+0.40) are the strongest predictors Full paper: arXiv:2601.00912 n Code & Data:GitHub The Problem: ChatGPT Visibility for Startups […]

Ver mais

arXiv:2602.02431v1 Announce Type: new Abstract: It is folklore that reusing training data more than once can improve the statistical efficiency of gradient-based learning. However, beyond linear regression, the theoretical advantage of full-batch gradient descent (GD, which always reuses all the data) over one-pass stochastic gradient descent (online SGD, which uses each data point only once) remains unclear. In this work, we consider learning a $d$-dimensional single-index model with a quadratic activation, for which it is known that one-pass […]

Ver mais

arXiv:2601.12186v1 Announce Type: new Abstract: Multi-domain thinking verifiers trained via Reinforcement Learning from Verifiable Rewards (RLVR) are a prominent fixture of the Large Language Model (LLM) post-training pipeline, owing to their ability to robustly rate and rerank model outputs. However, the adoption of such verifiers towards code generation has been comparatively sparse, with execution feedback constituting the dominant signal. Nonetheless, code verifiers remain valuable toward judging model outputs in scenarios where execution feedback is hard to obtain and […]

Ver mais