Search Results for author: Ehsan Shareghi

Found 60 papers, 38 papers with code

Integrating Transformers and Knowledge Graphs for Twitter Stance Detection

no code implementations WNUT (ACL) 2021 Thomas Clark, Costanza Conforti, Fangyu Liu, Zaiqiao Meng, Ehsan Shareghi, Nigel Collier

Stance detection (SD) entails classifying the sentiment of a text towards a given target, and is a relevant sub-task for opinion mining and social media analysis.

Knowledge Graphs Knowledge Probing +3

Fire Burns, Sword Cuts: Commonsense Inductive Bias for Exploration in Text-based Games

1 code implementation ACL 2022 Dongwon Ryu, Ehsan Shareghi, Meng Fang, Yunqiu Xu, Shirui Pan, Reza Haf

Text-based games (TGs) are exciting testbeds for developing deep reinforcement learning techniques due to their partially observed environments and large action spaces.

Deep Reinforcement Learning Efficient Exploration +4

On the Effect of Isotropy on VAE Representations of Text

1 code implementation ACL 2022 Lan Zhang, Wray Buntine, Ehsan Shareghi

Injecting desired geometric properties into text representations has attracted a lot of attention.

VerifiAgent: a Unified Verification Agent in Language Model Reasoning

no code implementations1 Apr 2025 Jiuzhou Han, Wray Buntine, Ehsan Shareghi

Large language models demonstrate remarkable reasoning capabilities but often produce unreliable or incorrect responses.

Language Modeling Language Modelling +1

SpeechDialogueFactory: Generating High-Quality Speech Dialogue Data to Accelerate Your Speech-LLM Development

1 code implementation31 Mar 2025 Minghan Wang, Ye Bai, Yuxia Wang, Thuy-Trang Vu, Ehsan Shareghi, Gholamreza Haffari

High-quality speech dialogue datasets are crucial for Speech-LLM development, yet existing acquisition methods face significant limitations.

Speech Synthesis Voice Cloning

ReasonGraph: Visualisation of Reasoning Paths

1 code implementation6 Mar 2025 Zongqian Li, Ehsan Shareghi, Nigel Collier

Large Language Models (LLMs) reasoning processes are challenging to analyze due to their complexity and the lack of organized visualization tools.

Assessing the Alignment of FOL Closeness Metrics with Human Judgement

1 code implementation15 Jan 2025 Ramya Keerthy Thatikonda, Wray Buntine, Ehsan Shareghi

Using ground-truth FOLs, we carefully designed various perturbations on the ground-truth to assess metric sensitivity.

Logical Reasoning Translation

One STEP at a time: Language Agents are Stepwise Planners

1 code implementation13 Nov 2024 Minh Nguyen, Ehsan Shareghi

We introduce STEP, a novel framework designed to efficiently learn from previous experiences to enhance the planning capabilities of language agents in future steps.

Audio Is the Achilles' Heel: Red Teaming Audio Large Multimodal Models

1 code implementation31 Oct 2024 Hao Yang, Lizhen Qu, Ehsan Shareghi, Gholamreza Haffari

Our results under these settings demonstrate that open-source audio LMMs suffer an average attack success rate of 69. 14% on harmful audio questions, and exhibit safety vulnerabilities when distracted with non-speech audio noise.

Red Teaming Safety Alignment

Jigsaw Puzzles: Splitting Harmful Questions to Jailbreak Large Language Models

1 code implementation15 Oct 2024 Hao Yang, Lizhen Qu, Ehsan Shareghi, Gholamreza Haffari

Moreover, JSP achieves a state-of-the-art attack success rate of 92% on GPT-4 on the harmful query benchmark, and exhibits strong resistant to defence strategies.

Strategies for Improving NL-to-FOL Translation with LLMs: Data Generation, Incremental Fine-Tuning, and Verification

1 code implementation24 Sep 2024 Ramya Keerthy Thatikonda, Jiuzhou Han, Wray Buntine, Ehsan Shareghi

Research in symbolic logical reasoning explored FOL generation using state-of-the-art LLMs (i. e., GPT-4) to produce FOL translations of natural language (NL) statements, but errors in translation are usually not the focus.

Data Augmentation Logical Reasoning +1

The Compressor-Retriever Architecture for Language Model OS

1 code implementation2 Sep 2024 Yuan Yang, Siheng Xiong, Ehsan Shareghi, Faramarz Fekri

Recent advancements in large language models (LLMs) have significantly enhanced their capacity to aggregate and process information across multiple modalities, enabling them to perform a wide range of tasks such as multimodal data querying, tool usage, web interactions, and handling long documents.

In-Context Learning Language Modeling +1

Towards Probing Speech-Specific Risks in Large Multimodal Models: A Taxonomy, Benchmark, and Insights

1 code implementation25 Jun 2024 Hao Yang, Lizhen Qu, Ehsan Shareghi, Gholamreza Haffari

Based on the taxonomy, we create a small-scale dataset for evaluating current LMMs capability in detecting these categories of risk.

Can LLMs Reason in the Wild with Programs?

1 code implementation19 Jun 2024 Yuan Yang, Siheng Xiong, Ali Payani, Ehsan Shareghi, Faramarz Fekri

To investigate this, we introduce the task of reasoning in the wild, where an LLM is tasked to solve a reasoning problem of unknown type by identifying the subproblems and their corresponding formalisms, and writing a program to solve each subproblem, guided by a tactic.

GSM8K Math

Exploring the Potential of Multimodal LLM with Knowledge-Intensive Multimodal ASR

1 code implementation16 Jun 2024 Minghan Wang, Yuxia Wang, Thuy-Trang Vu, Ehsan Shareghi, Gholamreza Haffari

Recent advancements in multimodal large language models (MLLMs) have made significant progress in integrating information across various modalities, yet real-world applications in educational and scientific domains remain challenging.

A Closer Look at Logical Reasoning with LLMs: The Choice of Tool Matters

1 code implementation1 Jun 2024 Long Hei Matthew Lam, Ramya Keerthy Thatikonda, Ehsan Shareghi

This paradigm has established the current state-of-the-art result in logical reasoning (i. e., deductive reasoning).

Logical Reasoning Translation

Aligning with Human Judgement: The Role of Pairwise Preference in Large Language Model Evaluators

1 code implementation25 Mar 2024 Yinhong Liu, Han Zhou, Zhijiang Guo, Ehsan Shareghi, Ivan Vulić, Anna Korhonen, Nigel Collier

Large Language Models (LLMs) have demonstrated promising capabilities as automatic evaluators in assessing the quality of generated natural language.

Language Modeling Language Modelling +1

Unlocking Structure Measuring: Introducing PDD, an Automatic Metric for Positional Discourse Coherence

1 code implementation15 Feb 2024 Yinhong Liu, Yixuan Su, Ehsan Shareghi, Nigel Collier

Recent large language models (LLMs) have shown remarkable performance in aligning generated text with user intentions across various tasks.

Coherence Evaluation Form +1

Equipping Language Models with Tool Use Capability for Tabular Data Analysis in Finance

no code implementations27 Jan 2024 Adrian Theuma, Ehsan Shareghi

Large language models (LLMs) have exhibited an array of reasoning capabilities but face challenges like error propagation and hallucination, particularly in specialised areas like finance, where data is heterogeneous, and precision is paramount.

Hallucination Language Modeling +2

Towards Uncertainty-Aware Language Agent

no code implementations25 Jan 2024 Jiuzhou Han, Wray Buntine, Ehsan Shareghi

We present the Uncertainty-Aware Language Agent (UALA), a framework that orchestrates the interaction between the agent and the external world using uncertainty quantification.

MMLU StrategyQA +1

Instruct-SCTG: Guiding Sequential Controlled Text Generation through Instructions

no code implementations19 Dec 2023 Yinhong Liu, Yixuan Su, Ehsan Shareghi, Nigel Collier

Instruction-tuned large language models have shown remarkable performance in aligning generated text with user intentions across various tasks.

Text Generation

POSQA: Probe the World Models of LLMs with Size Comparisons

1 code implementation20 Oct 2023 Chang Shu, Jiuzhou Han, Fangyu Liu, Ehsan Shareghi, Nigel Collier

Embodied language comprehension emphasizes that language understanding is not solely a matter of mental processing in the brain but also involves interactions with the physical and social environment.

Question Answering

FireAct: Toward Language Agent Fine-tuning

no code implementations9 Oct 2023 Baian Chen, Chang Shu, Ehsan Shareghi, Nigel Collier, Karthik Narasimhan, Shunyu Yao

Recent efforts have augmented language models (LMs) with external tools or environments, leading to the development of language agents that can reason and act.

Question Answering

Reward Engineering for Generating Semi-structured Explanation

1 code implementation15 Sep 2023 Jiuzhou Han, Wray Buntine, Ehsan Shareghi

Semi-structured explanation depicts the implicit process of a reasoner with an explicit representation.

Explanation Generation Reinforcement Learning (RL)

Investigating Pre-trained Audio Encoders in the Low-Resource Condition

1 code implementation28 May 2023 Hao Yang, Jinming Zhao, Gholamreza Haffari, Ehsan Shareghi

Pre-trained speech encoders have been central to pushing state-of-the-art results across various speech understanding and generation tasks.

Harnessing the Power of Large Language Models for Natural Language to First-Order Logic Translation

1 code implementation24 May 2023 Yuan Yang, Siheng Xiong, Ali Payani, Ehsan Shareghi, Faramarz Fekri

Translating natural language sentences to first-order logic (NL-FOL translation) is a longstanding challenge in the NLP and formal logic literature.

Formal Logic Sentence +1

PiVe: Prompting with Iterative Verification Improving Graph-based Generative Capability of LLMs

1 code implementation21 May 2023 Jiuzhou Han, Nigel Collier, Wray Buntine, Ehsan Shareghi

We show how a small language model could be trained to act as a verifier module for the output of an LLM~(i. e., ChatGPT, GPT-4), and to iteratively improve its performance via fine-grained corrective instructions.

Data Augmentation Graph Generation +3

Koala: An Index for Quantifying Overlaps with Pre-training Corpora

no code implementations26 Mar 2023 Thuy-Trang Vu, Xuanli He, Gholamreza Haffari, Ehsan Shareghi

In very recent years more attention has been placed on probing the role of pre-training data in Large Language Models (LLMs) downstream behaviour.

Memorization

Plug-and-Play Recipe Generation with Content Planning

no code implementations9 Dec 2022 Yinhong Liu, Yixuan Su, Ehsan Shareghi, Nigel Collier

Specifically, it optimizes the joint distribution of the natural language sequence and the global content plan in a plug-and-play manner.

Recipe Generation Sentence +1

Self-supervised Graph Masking Pre-training for Graph-to-Text Generation

1 code implementation19 Oct 2022 Jiuzhou Han, Ehsan Shareghi

Large-scale pre-trained language models (PLMs) have advanced Graph-to-Text (G2T) generation by processing the linearised version of a graph.

Decoder Text Generation

RedApt: An Adaptor for wav2vec 2 Encoding \\ Faster and Smaller Speech Translation without Quality Compromise

no code implementations16 Oct 2022 Jinming Zhao, Hao Yang, Gholamreza Haffari, Ehsan Shareghi

Pre-trained speech Transformers in speech translation (ST) have facilitated state-of-the-art (SotA) results; yet, using such encoders is computationally expensive.

Translation

Generating Synthetic Speech from SpokenVocab for Speech Translation

1 code implementation15 Oct 2022 Jinming Zhao, Gholamreza Haffar, Ehsan Shareghi

Training end-to-end speech translation (ST) systems requires sufficiently large-scale data, which is unavailable for most language pairs and domains.

Data Augmentation Machine Translation +3

On Reality and the Limits of Language Data: Aligning LLMs with Human Norms

no code implementations25 Aug 2022 Nigel H. Collier, Fangyu Liu, Ehsan Shareghi

Recent advancements in Large Language Models (LLMs) harness linguistic associations in vast natural language data for practical applications.

Common Sense Reasoning

M-Adapter: Modality Adaptation for End-to-End Speech-to-Text Translation

1 code implementation3 Jul 2022 Jinming Zhao, Hao Yang, Ehsan Shareghi, Gholamreza Haffari

End-to-end speech-to-text translation models are often initialized with pre-trained speech encoder and pre-trained text decoder.

Decoder Speech-to-Text +2

Rewire-then-Probe: A Contrastive Recipe for Probing Biomedical Knowledge of Pre-trained Language Models

1 code implementation ACL 2022 Zaiqiao Meng, Fangyu Liu, Ehsan Shareghi, Yixuan Su, Charlotte Collins, Nigel Collier

To catalyse the research in this direction, we release a well-curated biomedical knowledge probing benchmark, MedLAMA, which is constructed based on the Unified Medical Language System (UMLS) Metathesaurus.

Knowledge Probing Transfer Learning

The Neglected Sibling: Isotropic Gaussian Posterior for VAE

1 code implementation14 Oct 2021 Lan Zhang, Wray Buntine, Ehsan Shareghi

Deep generative models have been widely used in several areas of NLP, and various techniques have been proposed to augment them or address their training challenges.

Unsupervised Representation Disentanglement of Text: An Evaluation on Synthetic Datasets

1 code implementation ACL (RepL4NLP) 2021 Lan Zhang, Victor Prokhorov, Ehsan Shareghi

To highlight the challenges of achieving representation disentanglement for text domain in an unsupervised setting, in this paper we select a representative set of successfully applied models from the image domain.

Decoder Disentanglement +1

Combining Deep Generative Models and Multi-lingual Pretraining for Semi-supervised Document Classification

1 code implementation EACL 2021 Yi Zhu, Ehsan Shareghi, Yingzhen Li, Roi Reichart, Anna Korhonen

Semi-supervised learning through deep generative models and multi-lingual pretraining techniques have orchestrated tremendous success across different areas of NLP.

Classification Document Classification +1

A Closer Look at Few-Shot Crosslingual Transfer: The Choice of Shots Matters

no code implementations ACL 2021 Mengjie Zhao, Yi Zhu, Ehsan Shareghi, Ivan Vulić, Roi Reichart, Anna Korhonen, Hinrich Schütze

Few-shot crosslingual transfer has been shown to outperform its zero-shot counterpart with pretrained encoders like multilingual BERT.

Few-Shot Learning

Self-Alignment Pretraining for Biomedical Entity Representations

1 code implementation NAACL 2021 Fangyu Liu, Ehsan Shareghi, Zaiqiao Meng, Marco Basaldella, Nigel Collier

Despite the widespread success of self-supervised learning via masked language models (MLM), accurately capturing fine-grained semantic relationships in the biomedical domain remains a challenge.

Benchmarking Entity Linking +2

COMETA: A Corpus for Medical Entity Linking in the Social Media

1 code implementation EMNLP 2020 Marco Basaldella, Fangyu Liu, Ehsan Shareghi, Nigel Collier

Whilst there has been growing progress in Entity Linking (EL) for general language, existing datasets fail to address the complex nature of health terminology in layman's language.

Diversity Entity Linking

Learning Sparse Sentence Encoding without Supervision: An Exploration of Sparsity in Variational Autoencoders

1 code implementation ACL (RepL4NLP) 2021 Victor Prokhorov, Yingzhen Li, Ehsan Shareghi, Nigel Collier

It has been long known that sparsity is an effective inductive bias for learning efficient representation of data in vectors with fixed dimensionality, and it has been explored in many areas of representation learning.

Inductive Bias Representation Learning +3

On the Importance of the Kullback-Leibler Divergence Term in Variational Autoencoders for Text Generation

1 code implementation WS 2019 Victor Prokhorov, Ehsan Shareghi, Yingzhen Li, Mohammad Taher Pilehvar, Nigel Collier

While the explicit constraint naturally avoids posterior collapse, we use it to further understand the significance of the KL term in controlling the information transmitted through the VAE channel.

Text Generation

Bayesian Learning for Neural Dependency Parsing

no code implementations NAACL 2019 Ehsan Shareghi, Yingzhen Li, Yi Zhu, Roi Reichart, Anna Korhonen

While neural dependency parsers provide state-of-the-art accuracy for several languages, they still rely on large amounts of costly labeled training data.

Dependency Parsing parameter estimation +3

Structured Prediction of Sequences and Trees using Infinite Contexts

no code implementations9 Mar 2015 Ehsan Shareghi, Gholamreza Haffari, Trevor Cohn, Ann Nicholson

Linguistic structures exhibit a rich array of global phenomena, however commonly used Markov models are unable to adequately describe these phenomena due to their strong locality assumptions.

Part-Of-Speech Tagging Structured Prediction

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