Search Results for author: Zichen Liu

Found 14 papers, 8 papers with code

TEMP: Taxonomy Expansion with Dynamic Margin Loss through Taxonomy-Paths

1 code implementation EMNLP 2021 Zichen Liu, Hongyuan Xu, Yanlong Wen, Ning Jiang, Haiying Wu, Xiaojie Yuan

As an essential form of knowledge representation, taxonomies are widely used in various downstream natural language processing tasks.

Position Taxonomy Expansion

Selective Visual Prompting in Vision Mamba

1 code implementation12 Dec 2024 Yifeng Yao, Zichen Liu, Zhenyu Cui, Yuxin Peng, Jiahuan Zhou

To prevent the loss of discriminative information during state space propagation, SVP employs lightweight selective prompters for token-wise prompt generation, ensuring adaptive activation of the update and forget gates within Mamba blocks to promote discriminative information propagation.

Mamba State Space Models +1

Sample-Efficient Alignment for LLMs

1 code implementation3 Nov 2024 Zichen Liu, Changyu Chen, Chao Du, Wee Sun Lee, Min Lin

The results demonstrate that SEA achieves highly sample-efficient alignment with oracle's preferences, outperforming recent active exploration methods for LLMs.

Thompson Sampling

An Improved ESO-Based Line-of-Sight Guidance Law for Path Following of Underactuated Autonomous Underwater Helicopter With Nonlinear Tracking Differentiator and Anti-saturation Controller

no code implementations9 Oct 2024 Haoda Li, Zichen Liu, Jin Huang, Xinyu An, Ying Chen

By incorporating the nonlinear tracking differentiator and anti-saturation controller, the IELOS guidance law can precisely track sideslip angle and mitigate propeller thrust buffet compared to the classical Extended-state-observer based Line-of-Sight (ELOS) guidance law.

Bootstrapping Language Models with DPO Implicit Rewards

1 code implementation14 Jun 2024 Changyu Chen, Zichen Liu, Chao Du, Tianyu Pang, Qian Liu, Arunesh Sinha, Pradeep Varakantham, Min Lin

In this work, we make a novel observation that this implicit reward model can by itself be used in a bootstrapping fashion to further align the LLM.

Dynamic Typography: Bringing Text to Life via Video Diffusion Prior

no code implementations17 Apr 2024 Zichen Liu, Yihao Meng, Hao Ouyang, Yue Yu, Bolin Zhao, Daniel Cohen-Or, Huamin Qu

Through quantitative and qualitative evaluations, we demonstrate the effectiveness of our framework in generating coherent text animations that faithfully interpret user prompts while maintaining readability.

Vector Graphics

Locality Sensitive Sparse Encoding for Learning World Models Online

no code implementations23 Jan 2024 Zichen Liu, Chao Du, Wee Sun Lee, Min Lin

Unfortunately, NN-based models need re-training on all accumulated data at every interaction step to achieve FTL, which is computationally expensive for lifelong agents.

Continual Learning Model-based Reinforcement Learning

AnyHome: Open-Vocabulary Generation of Structured and Textured 3D Homes

no code implementations11 Dec 2023 Rao Fu, Zehao Wen, Zichen Liu, Srinath Sridhar

Inspired by cognitive theories, we introduce AnyHome, a framework that translates any text into well-structured and textured indoor scenes at a house-scale.

Diversity

TreeMAN: Tree-enhanced Multimodal Attention Network for ICD Coding

1 code implementation COLING 2022 Zichen Liu, Xuyuan Liu, Yanlong Wen, Guoqing Zhao, Fen Xia, Xiaojie Yuan

However, most previous works ignore the decisive information contained in structured medical data in EHRs, which is hard to be captured from the noisy clinical notes.

Adversarial Demonstration Attacks on Large Language Models

no code implementations24 May 2023 Jiongxiao Wang, Zichen Liu, Keun Hee Park, Zhuojun Jiang, Zhaoheng Zheng, Zhuofeng Wu, Muhao Chen, Chaowei Xiao

We propose a novel attack method named advICL, which aims to manipulate only the demonstration without changing the input to mislead the models.

In-Context Learning

Efficient Offline Policy Optimization with a Learned Model

1 code implementation12 Oct 2022 Zichen Liu, Siyi Li, Wee Sun Lee, Shuicheng Yan, Zhongwen Xu

Instead of planning with the expensive MCTS, we use the learned model to construct an advantage estimation based on a one-step rollout.

Offline RL

Cannot find the paper you are looking for? You can Submit a new open access paper.