Résultats de la recherche - decoding (algorithmics OR (algorithm OR (Algorithmssssics OR (algorithm OR algorithm))))

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

Table des matières: “… Feedback (RLHF) algorithms for aligning large language models (LLMs) with human preferences. However, DAAs…”
Accéder au texte intégral

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

Table des matières: “… ones, and beam search is commonly used to balance exploration and fluency during decoding. While deep…”
Accéder au texte intégral

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

Table des matières: “…, including: (1) encoder-only or decoder-only LLMs, (2) prompt-based parameter-efficient fine-tuning (PEFT…”
Accéder au texte intégral

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

Table des matières: “… an instruction-agnostic token compression algorithm based on a plug-and-play visual decoder to improve decoding…”
Accéder au texte intégral

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

Table des matières: “…, we introduce STARS, Synchronous Token Alignment for Robust Supervision, a decoding-time algorithm…”
Accéder au texte intégral

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

Table des matières: “… algorithms explore exponentially large Hilbert spaces. We recast qubit-efficient optimization as a geometric…”
Accéder au texte intégral

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

Table des matières: “… generation. Building on this, we introduce GoRL (Generative Online Reinforcement Learning), an algorithm…”
Accéder au texte intégral

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

Table des matières: “… with sampling strategies beyond greedy decoding. Our experiments validate that simple thresholding of the metric…”
Accéder au texte intégral

Hold Onto That Thought: Assessing KV Cache Compression On Reasoning par 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

Table des matières: “… attention computations, grows linearly with context length. A suite of compression algorithms has been…”
Accéder au texte intégral

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

Table des matières: “… inference. However, current confidence-driven decoding strategies are constrained by limited parallelism…”
Accéder au texte intégral

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

Table des matières: “…Standard autoregressive decoding in large language models (LLMs) is inherently short-sighted, often…”
Accéder au texte intégral

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

Table des matières: “… scaling (TTS) algorithms rely on autoregressive decoding, which is ill-suited to discrete diffusion…”
Accéder au texte intégral

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

Table des matières: “… RC, an iterative decoding algorithm that replaces standard autoregressive decoding during both…”
Accéder au texte intégral

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

Table des matières: “… Aware) decoder, a novel, training-free decoding algorithm that mitigates hallucinations at inference…”
Accéder au texte intégral

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

Table des matières: “… with the Viterbi decoding algorithm, replacing non-differentiable operations with smooth relaxations that enable…”
Accéder au texte intégral

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

Table des matières: “… and suboptimal discrete search algorithms or trial-and-error prompting to improve the online policy…”
Accéder au texte intégral

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

Table des matières: “… at inference time. Power sampling provides a principled way to bias decoding toward such modes by targeting…”
Accéder au texte intégral

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

Table des matières: “… preserving BF16 precision for decoding. By decoupling prefilling acceleration from decoding quality, Mix…”
Accéder au texte intégral

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

Table des matières: “… standard for decoding quantum low-density parity-check (QLDPC) codes. Recent advancements in this field…”
Accéder au texte intégral

NanoSpec: Accelerating Speculative Decoding using Minimalist In-Context Vocabularies par Chen, Zhiyang, Xu, Daliang, Zhang, Yinyuan, Wang, Chenghua, Xu, Mengwei, Ma, Yun

Table des matières: “… computational bottleneck on the final linear projection layer during speculative decoding. Existing vocabulary…”
Accéder au texte intégral