Yixin Liu

Yixin Liu

Auto-generated profile, based on publicly available information

Yixin Liu

Lehigh University

Research Focus & Metrics

Academic Career Length

6Years (2021-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
7 August 2024 in Nature Medicine
Verified
Journal Article
Closed access
PDF file
KZKai Zhang
Kai Zhang
RZRong Zhou
Rong Zhou
EAEashan Adhikarla
Eashan Adhikarla
ZYZhiling Yan
Zhiling Yan
YLYixin Liu
Yixin Liu
Lichao SunLichao Sun
Lichao Sun

22 January 2025 in ACM Computing Surveys
Verified
Journal Article
Open access
YCYihan Cao
Yihan Cao
SLSiyu Li
Siyu Li
YLYixin Liu
Yixin Liu
ZYZhiling Yan
Zhiling Yan
YDYutong Dai
Yutong Dai
Lichao SunLichao Sun
Lichao Sun
Recently, Artificial Intelligence Generated Content (AIGC) has gained significant attention from society, especially with the rise of Generative AI (GAI) techniques such as ChatGPT, GPT-4 [ 165 ], DALL-E-3 [ 184 ], and Sora [ 137 ]. AIGC involves using AI models to create digital content, such as images, music, and natural language, with the goal of making the content creation process more efficient and accessible. Large-scale models have become increasingly important in AIGC as they provide better intent extraction and generation results. This survey provides a comprehensive review of the history of generative models and recent advances in AIGC, focusing on both unimodal and multimodal interaction. From the perspective of unimodality, we introduce the generation tasks and relative models of text and image. From the perspective of multimodality, we introduce the cross-application between the modalities mentioned above. Finally, the survey discusses the existing open problems and future challenges in AIGC. Overall, this survey serves as a valuable resource for individuals interested in understanding the background and secrets behind the impressive performance of AIGC techniques.

11 June 2021 in IEEE Signal Processing Letters
Verified
Journal Article
Closed access
YLYixin Liu
Yixin Liu
YGYong Guo
Yong Guo
JGJiaxin Guo
Jiaxin Guo
LJLuoqian Jiang
Luoqian Jiang
JCJian Chen
Jian Chen
Channel pruning has become one of the predominant compression methods to deploy deep models on resource-constrained devices. Most channel pruning methods often use a fixed compression rate for all the layers of the model, which, however, may not be optimal. To address this issue, given a specific target compression rate, one can search for the optimal compression rate for each layer via some automated methods. Nevertheless, when we consider multiple compression rates, these methods have to repeat the channel pruning process multiple times, once for each rate, which can be unnecessary and inefficient. To tackle the problem, we propose a Conditional Automated Channel Pruning (CACP) method which simultaneously produces compressed models under different compression rates through a single channel pruning process. Specifically, CACP takes a set of compression rates and the original model as its input, and outputs the feasible compressed models that satisfy the considered compression rates. To learn CACP, we cast the layer-by-layer channel pruning process into a Markov decision process (MDP), in which we seek to solve a series of decision-making problems. Based on MDP, we develop a reinforcement learning (RL) framework with deep deterministic policy gradient (DDPG) to learn the optimal policy. To satisfy the constraint items in the optimization problem, we design a constraint-guaranteed method, which guides the agent to search for compressed models that satisfy the computational constraints by limiting the action space. Extensive experiments on CIFAR-10 and ImageNet datasets demonstrate the superiority of our method over existing methods.


4 February 2025 in IEEE Transactions on Knowledge and Data Engineering
Verified
Journal Article
Closed access
LHLijie Hu
Lijie Hu
XWXinhai Wang
Xinhai Wang
YLYixin Liu
Yixin Liu
NLNinghao Liu
Ninghao Liu
MHMengdi Huai
Mengdi Huai
DWDi Wang
Di Wang
Currently, attention mechanism has become a standard fixture in most state-of-the-art natural language processing (NLP) models, not only due to the outstanding performance it could gain but also due to plausible innate explanations for the behaviors of neural architectures it provides, which is notoriously difficult to analyze. However, recent studies show that attention is unstable against randomness and perturbations during training or testing, such as random seeds and slight perturbation of embedding vectors, which impedes it from becoming a faithful explanation tool. Thus, a natural question is whether we can find some substitute for the current attention that is more stable and could keep the most important characteristics of explanation and prediction of attention. In this paper, to resolve the problem, we provide a rigorous definition of such alternate namely SEAT (Stable and Explainable Attention). Specifically, a SEAT should have the following three properties: (1) Its prediction distribution is enforced to be close to the distribution based on the vanilla attention; (2) Its top-$k$ indices have large overlaps with those of the vanilla attention; (3) It is robust w.r.t perturbations, i.e., any slight perturbation on SEAT will not change the prediction distribution too much, which implicitly indicates that it is stable to randomness and perturbations. To further improve the interpretability stability against perturbations, based on SEAT we provide another definition called SEAT++. Then we propose a method to get a SEAT++, which could be considered an ad hoc modification for canonical attention. Finally, through intensive experiments on various datasets, we compare our SEAT and SEAT++ with other baseline methods using RNN, BiLSTM, and BERT architectures via six different evaluation metrics for model interpretation, stability, and accuracy. Results show that SEAT and SEAT++ are more stable against different perturbations and randomness while also keeping the explainability of attention, which indicates they provide more faithful explanations. Moreover, compared with vanilla attention, there is almost no utility (accuracy) degradation for SEAT and SEAT++.

30 July 2025 in Harvard Data Science Review
Verified
Journal Article
Open access
PDF file
XLXiaokang Liu
Xiaokang Liu
JYJun Yu
Jun Yu
YDYutong Dai
Yutong Dai
YSYishan Shen
Yishan Shen
JCJianmin Chen
Jianmin Chen
YCYong Chen
Yong Chen
In Part I of this survey, we introduced the fundamentals of multitask learning (MTL), including its definition, significance, and core operational mechanisms. This Part II builds on that foundation by summarizing existing knowledge and reviewing key methodologies that underpin MTL, with a focus on regularization and optimization techniques that facilitate effective information sharing across tasks. We begin by exploring a variety of regularization approaches, including sparsity-inducing feature selection methods, low-rank methods, weight matrix decomposition techniques, prior-sharing strategies, and task clustering methods, each offering a unique lens through which to capture task similarities and disentangle shared from task-specific structures. Next, we delve into optimization strategies, including scalarization approaches, multiobjective optimization methods, adversarial training, and neural architecture search, which are used to align gradient directions across tasks and improve convergence. For each category of methods, we present key formulations and technical details of representative works, concluding with a summary of insights and practical takeaways. By presenting existing methods and summarizing their strengths and limitations, this part of the survey aims to provide conceptual insights that may inspire the broader application of MTL and inform future research directions.“Multi-Task Learning 1997–2024” is a three-part article. Part I, “Fundamentals” can be read here . Part III, “Applications” can be read here .