Risultati della ricerca - decoding (algorithm OR (algorithmics OR Algorithmssics))

An Exploratory Framework for Future SETI Applications: Detecting Generative Reactivity via Language Models di Yu, Po-Chieh

Sommario: “... responses in language models. Instead of assuming that extraterrestrial signals must be decoded, we evaluate...”
Testo

Towards Bridging the Reward-Generation Gap in Direct Alignment Algorithms di Xiao, Zeguan, Chen, Yun, Chen, Guanhua, Tang, Ke

Sommario: “... Feedback (RLHF) algorithms for aligning large language models (LLMs) with human preferences. However, DAAs...”
Testo

ABS: Enforcing Constraint Satisfaction On Generated Sequences Via Automata-Guided Beam Search di Collura, Vincenzo, Tit, Karim, Bussi, Laura, Giunchiglia, Eleonora, Cordy, Maxime

Sommario: “... ones, and beam search is commonly used to balance exploration and fluency during decoding. While deep...”
Testo

A Study of Large Language Models for Patient Information Extraction: Model Architecture, Fine-Tuning Strategy, and Multi-task Instruction Tuning di Peng, Cheng, Dong, Xinyu, Lyu, Mengxian, Paredes, Daniel, Zhang, Yaoyun, Wu, Yonghui

Sommario: “..., including: (1) encoder-only or decoder-only LLMs, (2) prompt-based parameter-efficient fine-tuning (PEFT...”
Testo

Reading Images Like Texts: Sequential Image Understanding in Vision-Language Models di Li, Yueyan, Zhao, Chenggong, Zang, Zeyuan, Yuan, Caixia, Wang, Xiaojie

Sommario: “... an instruction-agnostic token compression algorithm based on a plug-and-play visual decoder to improve decoding...”
Testo

STARS: Synchronous Token Alignment for Robust Supervision in Large Language Models di Quamar, Mohammad Atif, Areeb, Mohammad, Kuznetsov, Mikhail, Ozmen, Muslum Ozgur, Celik, Z. Berkay

Sommario: “..., we introduce STARS, Synchronous Token Alignment for Robust Supervision, a decoding-time algorithm...”
Testo

An Information-Minimal Geometry for Qubit-Efficient Optimization di Ma, Gordon, Angelakis, Dimitris G.

Sommario: “... algorithms explore exponentially large Hilbert spaces. We recast qubit-efficient optimization as a geometric...”
Testo

Evolving Diffusion and Flow Matching Policies for Online Reinforcement Learning di Zhang, Chubin, Wan, Zhenglin, Chen, Feng, Yang, Fuchao, Feng, Lang, Zhou, Yaxin, Yu, Xingrui, You, Yang, Tsang, Ivor, An, Bo

Sommario: “... generation. Building on this, we introduce GoRL (Generative Online Reinforcement Learning), an algorithm...”
Testo

Short-Context Dominance: How Much Local Context Natural Language Actually Needs? di Vakilian, Vala, Wang, Zimeng, Rawat, Ankit Singh, Thrampoulidis, Christos

Sommario: “... with sampling strategies beyond greedy decoding. Our experiments validate that simple thresholding of the metric...”
Testo

Hold Onto That Thought: Assessing KV Cache Compression On Reasoning di Liu, Minghui, Palnitkar, Aadi, Rabbani, Tahseen, Jae, Hyunwoo, Sang, Kyle Rui, Yao, Dixi, Shabihi, Shayan, Zhao, Fuheng, Li, Tian, Zhang, Ce, Huang, Furong, Zhang, Kunpeng

Sommario: “... attention computations, grows linearly with context length. A suite of compression algorithms has been...”
Testo

LoPA: Scaling dLLM Inference via Lookahead Parallel Decoding di Xu, Chenkai, Jin, Yijie, Li, Jiajun, Tu, Yi, Long, Guoping, Tu, Dandan, Song, Mingcong, Si, Hongjie, Hou, Tianqi, Yan, Junchi, Deng, Zhijie

Sommario: “... inference. However, current confidence-driven decoding strategies are constrained by limited parallelism...”
Testo

Martingale Foresight Sampling: A Principled Approach to Inference-Time LLM Decoding di Li, Huayu, He, ZhengXiao, Tian, Siyuan, Wen, Jinghao, Li, Ao

Sommario: “...Standard autoregressive decoding in large language models (LLMs) is inherently short-sighted, often...”
Testo

Prism: Efficient Test-Time Scaling via Hierarchical Search and Self-Verification for Discrete Diffusion Language Models di Bai, Jinbin, Li, Yixuan, Zhu, Yuchen, Xin, Yi, Shi, Qingyu, Feng, Aosong, Liu, Xiaohong, Tao, Molei, Xue, Jianru, Li, Xiangtai, Yang, Ming-Hsuan

Sommario: “... scaling (TTS) algorithms rely on autoregressive decoding, which is ill-suited to discrete diffusion...”
Testo

Reasoning Cache: Continual Improvement Over Long Horizons via Short-Horizon RL di Wu, Ian, Qu, Yuxiao, Setlur, Amrith, Kumar, Aviral

Sommario: “... RC, an iterative decoding algorithm that replaces standard autoregressive decoding during both...”
Testo

Listen to the Layers: Mitigating Hallucinations with Inter-Layer Disagreement di Subbalakshmi, Koduvayur, Ujjal, Sabbir Hossain, Mangichetty, Venkata Krishna Teja, Soofi, Nastaran Jamalipour

Sommario: “... Aware) decoder, a novel, training-free decoding algorithm that mitigates hallucinations at inference...”
Testo

ViterbiPlanNet: Injecting Procedural Knowledge via Differentiable Viterbi for Planning in Instructional Videos di Seminara, Luigi, Moltisanti, Davide, Furnari, Antonino

Sommario: “... with the Viterbi decoding algorithm, replacing non-differentiable operations with smooth relaxations that enable...”
Testo

$\nabla$-Reasoner: LLM Reasoning via Test-Time Gradient Descent in Latent Space di Wang, Peihao, Cai, Ruisi, Wang, Zhen, Mei, Hongyuan, Liu, Qiang, Li, Pan, Wang, Zhangyang

Sommario: “... and suboptimal discrete search algorithms or trial-and-error prompting to improve the online policy...”
Testo

The Model Knows, the Decoder Finds: Future Value Guided Particle Power Sampling di Nguyen, Tu, Zimmer, Matthieu, Tutunov, Rasul, Ji, Xiaotong, Ammar, Haitham Bou

Sommario: “... at inference time. Power sampling provides a principled way to bias decoding toward such modes by targeting...”
Testo

Mix-Quant: Quantized Prefilling, Precise Decoding for Agentic LLMs di Lu, Haiquan, Chen, Zigeng, Fang, Gongfan, Ma, Xinyin, Wang, Xinchao

Sommario: “... preserving BF16 precision for decoding. By decoupling prefilling acceleration from decoding quality, Mix...”
Testo

Best-First Ordered Statistics Decoding of Quantum LDPC Codes di Banfi, Michele, Ferrari, Marco, Favano, Antonino, Tarable, Alberto, Barletta, Luca

Sommario: “... standard for decoding quantum low-density parity-check (QLDPC) codes. Recent advancements in this field...”
Testo