Search Results for author: Jiaqing Liang

Found 38 papers, 16 papers with code

Tokenization Matters! Degrading Large Language Models through Challenging Their Tokenization

no code implementations27 May 2024 Dixuan Wang, Yanda Li, Junyuan Jiang, Zepeng Ding, Guochao Jiang, Jiaqing Liang, Deqing Yang

Our empirical results reveal that our ADT is highly effective on challenging the tokenization of leading LLMs, including GPT-4o, Llama-3, Qwen2. 5-max and so on, thus degrading these LLMs' capabilities.

SED: Self-Evaluation Decoding Enhances Large Language Models for Better Generation

no code implementations26 May 2024 Ziqin Luo, Haixia Han, Haokun Zhao, Guochao Jiang, Chengyu Du, Tingyun Li, Jiaqing Liang, Deqing Yang, Yanghua Xiao

Existing Large Language Models (LLMs) generate text through unidirectional autoregressive decoding methods to respond to various user queries.

Decision Making

From Complex to Simple: Enhancing Multi-Constraint Complex Instruction Following Ability of Large Language Models

1 code implementation24 Apr 2024 Qianyu He, Jie Zeng, Qianxi He, Jiaqing Liang, Yanghua Xiao

It is imperative for Large language models (LLMs) to follow instructions with elaborate requirements (i. e. Complex Instructions Following).

Instruction Following

AutoCrawler: A Progressive Understanding Web Agent for Web Crawler Generation

1 code implementation19 Apr 2024 Wenhao Huang, Chenghao Peng, Zhixu Li, Jiaqing Liang, Yanghua Xiao, Liqian Wen, Zulong Chen

We propose AutoCrawler, a two-stage framework that leverages the hierarchical structure of HTML for progressive understanding.

Action Generation

Enhancing Confidence Expression in Large Language Models Through Learning from Past Experience

no code implementations16 Apr 2024 Haixia Han, Tingyun Li, Shisong Chen, Jie Shi, Chengyu Du, Yanghua Xiao, Jiaqing Liang, Xin Lin

Specifically, we first identify three key problems: (1) How to capture the inherent confidence of the LLM?

ToNER: Type-oriented Named Entity Recognition with Generative Language Model

1 code implementation14 Apr 2024 Guochao Jiang, Ziqin Luo, Yuchen Shi, Dixuan Wang, Jiaqing Liang, Deqing Yang

In recent years, the fine-tuned generative models have been proven more powerful than the previous tagging-based or span-based models on named entity recognition (NER) task.

Binary Classification Language Modelling +4

Large Language Model Can Continue Evolving From Mistakes

no code implementations11 Apr 2024 Haokun Zhao, Haixia Han, Jie Shi, Chengyu Du, Jiaqing Liang, Yanghua Xiao

As world knowledge evolves and new task paradigms emerge, Large Language Models (LLMs) often fall short of meeting new demands due to knowledge deficiencies and outdated information.

Continual Learning Continual Pretraining +4

Reason from Fallacy: Enhancing Large Language Models' Logical Reasoning through Logical Fallacy Understanding

no code implementations4 Apr 2024 Yanda Li, Dixuan Wang, Jiaqing Liang, Guochao Jiang, Qianyu He, Yanghua Xiao, Deqing Yang

Large Language Models (LLMs) have demonstrated good performance in many reasoning tasks, but they still struggle with some complicated reasoning tasks including logical reasoning.

Logical Fallacies Logical Reasoning

Small Language Model Can Self-correct

no code implementations14 Jan 2024 Haixia Han, Jiaqing Liang, Jie Shi, Qianyu He, Yanghua Xiao

In this paper, we introduce the \underline{I}ntrinsic \underline{S}elf-\underline{C}orrection (ISC) in generative language models, aiming to correct the initial output of LMs in a self-triggered manner, even for those small LMs with 6 billion parameters.

Language Modelling

ConcEPT: Concept-Enhanced Pre-Training for Language Models

no code implementations11 Jan 2024 Xintao Wang, Zhouhong Gu, Jiaqing Liang, Dakuan Lu, Yanghua Xiao, Wei Wang

In this paper, we propose ConcEPT, which stands for Concept-Enhanced Pre-Training for language models, to infuse conceptual knowledge into PLMs.

Entity Linking Entity Typing

Enhancing Quantitative Reasoning Skills of Large Language Models through Dimension Perception

no code implementations29 Dec 2023 Yuncheng Huang, Qianyu He, Jiaqing Liang, Sihang Jiang, Yanghua Xiao, Yunwen Chen

Hence, we present a framework to enhance the quantitative reasoning ability of language models based on dimension perception.

Towards Visual Taxonomy Expansion

1 code implementation12 Sep 2023 Tinghui Zhu, Jingping Liu, Jiaqing Liang, Haiyun Jiang, Yanghua Xiao, ZongYu Wang, Rui Xie, Yunsen Xian

Specifically, on the Chinese taxonomy dataset, our method significantly improves accuracy by 8. 75 %.

Taxonomy Expansion

KnowledGPT: Enhancing Large Language Models with Retrieval and Storage Access on Knowledge Bases

no code implementations17 Aug 2023 Xintao Wang, Qianwen Yang, Yongting Qiu, Jiaqing Liang, Qianyu He, Zhouhong Gu, Yanghua Xiao, Wei Wang

Large language models (LLMs) have demonstrated impressive impact in the field of natural language processing, but they still struggle with several issues regarding, such as completeness, timeliness, faithfulness and adaptability.

Retrieval World Knowledge

Causality-aware Concept Extraction based on Knowledge-guided Prompting

1 code implementation3 May 2023 Siyu Yuan, Deqing Yang, Jinxi Liu, Shuyu Tian, Jiaqing Liang, Yanghua Xiao, Rui Xie

The prompt adopts the topic of the given entity from the existing knowledge in KGs to mitigate the spurious co-occurrence correlations between entities and biased concepts.

Knowledge Graphs Natural Language Understanding

GANTEE: Generative Adversatial Network for Taxonomy Entering Evaluation

no code implementations25 Mar 2023 Zhouhong Gu, Sihang Jiang, Jingping Liu, Yanghua Xiao, Hongwei Feng, Zhixu Li, Jiaqing Liang, Jian Zhong

The previous methods suffer from low-efficiency since they waste much time when most of the new coming concepts are indeed noisy concepts.

Generative Adversarial Network Taxonomy Expansion

MAPS-KB: A Million-scale Probabilistic Simile Knowledge Base

2 code implementations10 Dec 2022 Qianyu He, Xintao Wang, Jiaqing Liang, Yanghua Xiao

The ability to understand and generate similes is an imperative step to realize human-level AI.

Generative Entity Typing with Curriculum Learning

1 code implementation6 Oct 2022 Siyu Yuan, Deqing Yang, Jiaqing Liang, Zhixu Li, Jinxi Liu, Jingyue Huang, Yanghua Xiao

To overcome these drawbacks, we propose a novel generative entity typing (GET) paradigm: given a text with an entity mention, the multiple types for the role that the entity plays in the text are generated with a pre-trained language model (PLM).

Entity Typing Language Modelling

Large-scale Multi-granular Concept Extraction Based on Machine Reading Comprehension

1 code implementation30 Aug 2022 Siyu Yuan, Deqing Yang, Jiaqing Liang, Jilun Sun, Jingyue Huang, Kaiyan Cao, Yanghua Xiao, Rui Xie

In order to supply existing KGs with more fine-grained and new concepts, we propose a novel concept extraction framework, namely MRC-CE, to extract large-scale multi-granular concepts from the descriptive texts of entities.

Descriptive Knowledge Graphs +1

FL-Tuning: Layer Tuning for Feed-Forward Network in Transformer

1 code implementation30 Jun 2022 Jingping Liu, Yuqiu Song, Kui Xue, Hongli Sun, Chao Wang, Lihan Chen, Haiyun Jiang, Jiaqing Liang, Tong Ruan

Specifically, we focus on layer tuning for feed-forward network in the Transformer, namely FL-tuning.

Model Optimization

Language Models as Knowledge Embeddings

1 code implementation25 Jun 2022 Xintao Wang, Qianyu He, Jiaqing Liang, Yanghua Xiao

In this paper, we propose LMKE, which adopts Language Models to derive Knowledge Embeddings, aiming at both enriching representations of long-tail entities and solving problems of prior description-based methods.

Contrastive Learning Link Prediction +1

Tackling Math Word Problems with Fine-to-Coarse Abstracting and Reasoning

no code implementations17 May 2022 Ailisi Li, Xueyao Jiang, Bang Liu, Jiaqing Liang, Yanghua Xiao

Math Word Problems (MWP) is an important task that requires the ability of understanding and reasoning over mathematical text.

Math

Rule Mining over Knowledge Graphs via Reinforcement Learning

no code implementations21 Feb 2022 Lihan Chen, Sihang Jiang, Jingping Liu, Chao Wang, Sheng Zhang, Chenhao Xie, Jiaqing Liang, Yanghua Xiao, Rui Song

Knowledge graphs (KGs) are an important source repository for a wide range of applications and rule mining from KGs recently attracts wide research interest in the KG-related research community.

Knowledge Graphs reinforcement-learning +1

Semantic-based Data Augmentation for Math Word Problems

no code implementations7 Jan 2022 Ailisi Li, Jiaqing Liang, Yanghua Xiao

In this paper, we propose a set of novel data augmentation approaches to supplement existing datasets with such data that are augmented with different kinds of local variances, and help to improve the generalization ability of current neural models.

Data Augmentation Math

Collective Loss Function for Positive and Unlabeled Learning

no code implementations6 May 2020 Chenhao Xie, Qiao Cheng, Jiaqing Liang, Lihan Chen, Yanghua Xiao

On the contrary, traditional machine learning algorithms often rely on negative examples, otherwise the model would be prone to collapse and always-true predictions.

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