東華大學圖書館 |

語系: 繁體中文

說明(常見問題)

回圖書館首頁

手機版館藏查詢

登入

回首頁 到查詢結果 [ null ]

切換: 標籤 | MARC模式 | ISBD

Towards Robust and Generalizable Mac...

Wang, Yihan,

FindBook

Google Book

Amazon

博客來

Towards Robust and Generalizable Machine Learning Models /

紀錄類型:	書目-電子資源 : Monograph/item
正題名/作者:	Towards Robust and Generalizable Machine Learning Models // Yihan Wang.
作者:	Wang, Yihan,
面頁冊數:	1 electronic resource (227 pages)
附註:	Source: Dissertations Abstracts International, Volume: 86-11, Section: B.
Contained By:	Dissertations Abstracts International86-11B.
標題:	Computer science. -
電子資源:	https://pqdd.sinica.edu.tw/twdaoapp/servlet/advanced?query=31997400
ISBN:	9798314887936

Towards Robust and Generalizable Machine Learning Models /
Wang, Yihan,

Towards Robust and Generalizable Machine Learning Models /Yihan Wang. - 1 electronic resource (227 pages)

Source: Dissertations Abstracts International, Volume: 86-11, Section: B.

With the rapid advancement and widespread adoption of machine learning technologies, concerns about model robustness and generalization have become increasingly significant. This dissertation discusses and addresses critical challenges in developing machine learning models that are both robust against adversarial examples and capable of generalizing across distribution shifts. We first investigate certified robustness methods for neural networks, proposing Fast-IBP, a novel approach that achieves state-of-the-art certified robust accuracy with significantly reduced training time through improved initialization and normalization techniques. We also provide theoretical analysis on the limitations of width scaling in Interval Bound Propagation training. Moving beyond traditional robustness concerns, we explore challenges in responsible deployment of large language models (LLMs), developing comprehensive red-teaming tests for popular LLM text detection methods and proposing a training-free backtranslation defense against jailbreaking attacks. Finally, we extend our discussions from robustness in the deployment stage to the training stage. We provide theoretical analysis towards the limitations of prompt-tuning, identifying its representation power and failure cases. We also identify format overfitting as a partial explanation for language model overfitting when fine-tuned on specific downstream tasks, and introduce PROMOT, a two-stage fine-tuning strategy that mitigates over-specialization while maintaining or improving in-context learning performance on unfinetuned tasks. Moving from single-task fine-tuning to general instruction fine-tuning, we identify the loss of context awareness after instruction finetuning for language models. Based on empirical observations, we propose a method to identify context-dependent training examples and mitigate the performance loss. Our work contributes significant advancements in building machine learning systems that are not only powerful but also reliable, safe, and aligned with human values.

English

ISBN: 9798314887936Subjects--Topical Terms:

523869
Computer science.
Subjects--Index Terms:

Large language models

Towards Robust and Generalizable Machine Learning Models /
LDR:03485nmm a22004333i 4500 001 2432304
005 20250707081340.5
006 m o d
007 cr|nu||||||||
008 261011s2025 miu||||||m |||||||eng d
020 $a 9798314887936
035 $a (MiAaPQD)AAI31997400
035 $a AAI31997400
040 $a MiAaPQD $b eng $c MiAaPQD $e rda
100 1 $a Wang, Yihan, $e author. $3 3818551
245 1 0 $a Towards Robust and Generalizable Machine Learning Models / $c Yihan Wang.
264 1 $a Ann Arbor : $b ProQuest Dissertations & Theses, $c 2025
300 $a 1 electronic resource (227 pages)
336 $a text $b txt $2 rdacontent
337 $a computer $b c $2 rdamedia
338 $a online resource $b cr $2 rdacarrier
500 $a Source: Dissertations Abstracts International, Volume: 86-11, Section: B.
500 $a Advisors: Hsieh, Cho-Jui Committee members: Chang, Kai-Wei; Peng, Nanyun; Gu, Quanquan.
502 $b Ph.D. $c University of California, Los Angeles $d 2025.
520 $a With the rapid advancement and widespread adoption of machine learning technologies, concerns about model robustness and generalization have become increasingly significant. This dissertation discusses and addresses critical challenges in developing machine learning models that are both robust against adversarial examples and capable of generalizing across distribution shifts. We first investigate certified robustness methods for neural networks, proposing Fast-IBP, a novel approach that achieves state-of-the-art certified robust accuracy with significantly reduced training time through improved initialization and normalization techniques. We also provide theoretical analysis on the limitations of width scaling in Interval Bound Propagation training. Moving beyond traditional robustness concerns, we explore challenges in responsible deployment of large language models (LLMs), developing comprehensive red-teaming tests for popular LLM text detection methods and proposing a training-free backtranslation defense against jailbreaking attacks. Finally, we extend our discussions from robustness in the deployment stage to the training stage. We provide theoretical analysis towards the limitations of prompt-tuning, identifying its representation power and failure cases. We also identify format overfitting as a partial explanation for language model overfitting when fine-tuned on specific downstream tasks, and introduce PROMOT, a two-stage fine-tuning strategy that mitigates over-specialization while maintaining or improving in-context learning performance on unfinetuned tasks. Moving from single-task fine-tuning to general instruction fine-tuning, we identify the loss of context awareness after instruction finetuning for language models. Based on empirical observations, we propose a method to identify context-dependent training examples and mitigate the performance loss. Our work contributes significant advancements in building machine learning systems that are not only powerful but also reliable, safe, and aligned with human values.
546 $a English
590 $a School code: 0031
650 4 $a Computer science. $3 523869
650 4 $a Engineering. $3 586835
650 4 $a Information technology. $3 532993
653 $a Large language models
653 $a Machine learning
653 $a Robust accuracy
653 $a Jailbreaking attacks
690 $a 0984
690 $a 0489
690 $a 0800
690 $a 0537
710 2 $a University of California, Los Angeles. $b Computer Science 0201. $e degree granting institution. $3 3816290
720 1 $a Hsieh, Cho-Jui $e degree supervisor.
773 0 $t Dissertations Abstracts International $g 86-11B.
790 $a 0031
791 $a Ph.D.
792 $a 2025
856 4 0 $u https://pqdd.sinica.edu.tw/twdaoapp/servlet/advanced?query=31997400