English
Related papers

Related papers: Privileged Information Distillation for Language M…

200 papers

On-policy distillation (OPD) trains a student on its own rollouts with token-level teacher supervision. Recent selective OPD methods exploit the non-uniformity of OPD signals by prioritizing high-entropy or high-disagreement tokens. We…

Machine Learning · Computer Science 2026-05-27 Yuanyi Wang , Su Lu , Yanggan Gu , Pengkai Wang , Yifan Yang , Zhaoyi Yan , Congkai Xie , Jianmin Wu , Hongxia Yang

Recent advances in knowledge distillation (KD) have enabled smaller student models to approach the performance of larger teacher models. However, popular methods such as supervised KD and on-policy KD, are adversely impacted by the…

Computation and Language · Computer Science 2025-04-29 Wenda Xu , Rujun Han , Zifeng Wang , Long T. Le , Dhruv Madeka , Lei Li , William Yang Wang , Rishabh Agarwal , Chen-Yu Lee , Tomas Pfister

Distillation efforts have led to language models that are more compact and efficient without serious drops in performance. The standard approach to distillation trains a student model against two objectives: a task-specific objective (e.g.,…

Computation and Language · Computer Science 2022-06-07 Zhengxuan Wu , Atticus Geiger , Josh Rozner , Elisa Kreiss , Hanson Lu , Thomas Icard , Christopher Potts , Noah D. Goodman

On-policy distillation (OPD) trains student models under their own induced distribution while leveraging supervision from stronger teachers. We identify a failure mode of OPD: as training progresses, on-policy rollouts can undergo abrupt…

Computation and Language · Computer Science 2026-04-10 Feng Luo , Yu-Neng Chuang , Guanchu Wang , Zicheng Xu , Xiaotian Han , Tianyi Zhang , Vladimir Braverman

Language model (LM) post-training (or alignment) involves maximizing a reward function that is derived from preference annotations. Direct Preference Optimization (DPO) is a popular offline alignment method that trains a policy directly on…

Machine Learning · Computer Science 2025-03-04 Adam Fisch , Jacob Eisenstein , Vicky Zayats , Alekh Agarwal , Ahmad Beirami , Chirag Nagpal , Pete Shaw , Jonathan Berant

Leveraging privileged information (PI), or features available during training but not at test time, has recently been shown to be an effective method for addressing label noise. However, the reasons for its effectiveness are not well…

Partial observability of the underlying states generally presents significant challenges for reinforcement learning (RL). In practice, certain \emph{privileged information}, e.g., the access to states from simulators, has been exploited in…

Machine Learning · Computer Science 2025-02-24 Yang Cai , Xiangyu Liu , Argyris Oikonomou , Kaiqing Zhang

Pre-trained language models such as BERT have proven to be highly effective for natural language processing (NLP) tasks. However, the high demand for computing resources in training such models hinders their application in practice. In…

Computation and Language · Computer Science 2019-08-27 Siqi Sun , Yu Cheng , Zhe Gan , Jingjing Liu

On-policy distillation (OPD) has shown strong potential for transferring reasoning ability from frontier or domain-specific models to smaller students. While effective on static single-turn tasks, its behavior in multi-turn agent settings…

Machine Learning · Computer Science 2026-04-30 Jiaqi Wang , Wenhao Zhang , Weijie Shi , Yaliang Li , James Cheng

We present a novel approach to knowledge transfer in model-based reinforcement learning, addressing the critical challenge of deploying large world models in resource-constrained environments. Our method efficiently distills a high-capacity…

Machine Learning · Computer Science 2025-07-03 Dmytro Kuzmenko , Nadiya Shvai

Reinforcement learning from verifiable rewards (RLVR) suffers from sparse outcome signals, creating severe exploration bottlenecks on complex reasoning tasks. Recent on-policy self-distillation methods attempt to address this by utilizing…

Machine Learning · Computer Science 2026-05-20 Yang Li , Erik Nijkamp , Semih Yavuz , Shafiq Joty

Knowledge distillation has proven effective for model compression by transferring knowledge from a larger network called the teacher to a smaller network called the student. Current knowledge distillation in time series is predominantly…

Machine Learning · Computer Science 2026-01-12 Nilushika Udayangani Hewa Dehigahawattage , Kishor Nandakishor , Marimuthu Palaniswami

Knowledge distillation is used, in generative language modeling, to train a smaller student model using the help of a larger teacher model, resulting in improved capabilities for the student model. In this paper, we formulate a more general…

Computation and Language · Computer Science 2025-02-26 Guanlin Liu , Anand Ramachandran , Tanmay Gangwani , Yan Fu , Abhinav Sethy

Vision-based deep reinforcement learning (RL) typically obtains performance benefit by using high capacity and relatively large convolutional neural networks (CNN). However, a large network leads to higher inference costs (power, latency,…

Machine Learning · Computer Science 2019-05-01 Sam Green , Craig M. Vineyard , Çetin Kaya Koç

Knowledge distillation (KD) is one of the prominent techniques for model compression. In this method, the knowledge of a large network (teacher) is distilled into a model (student) with usually significantly fewer parameters. KD tries to…

Machine Learning · Computer Science 2023-01-31 Aref Jafari , Mehdi Rezagholizadeh , Ali Ghodsi

Diffusion models generate high-quality images through progressive denoising but are computationally intensive due to large model sizes and repeated sampling. Knowledge distillation, which transfers knowledge from a complex teacher to a…

Machine Learning · Computer Science 2025-04-04 Dohyun Kim , Sehwan Park , Geonhee Han , Seung Wook Kim , Paul Hongsuck Seo

Causal language models have demonstrated remarkable capabilities, but their size poses significant challenges for deployment in resource-constrained environments. Knowledge distillation, a widely-used technique for transferring knowledge…

Machine Learning · Computer Science 2025-03-03 Makoto Shing , Kou Misaki , Han Bao , Sho Yokoi , Takuya Akiba

On-policy distillation (OPD) has become a promising paradigm for reasoning-oriented post-training of large language models (LLMs), especially when combined with reinforcement learning from verifiable rewards (RLVR). Existing OPD methods…

Personalized speech enhancement (PSE) models achieve promising results compared with unconditional speech enhancement models due to their ability to remove interfering speech in addition to background noise. Unlike unconditional speech…

Audio and Speech Processing · Electrical Eng. & Systems 2022-11-08 Hassan Taherian , Sefik Emre Eskimez , Takuya Yoshioka

In self-supervised learning, self-distilled methods have shown impressive performance, learning representations useful for downstream tasks and even displaying emergent properties. However, state-of-the-art methods usually rely on ensembles…

Machine Learning · Computer Science 2026-05-01 Esteban Rodríguez-Betancourt , Edgar Casasola-Murillo
‹ Prev 1 4 5 6 7 8 10 Next ›