arXiv subjects

Bang Liu

Publications and source records attributed to Bang Liu.

At least 109 records · Page 6Linked to original sources

arXiv2024

OPEx: A Component-Wise Analysis of LLM-Centric Agents in Embodied Instruction Following

Embodied Instruction Following (EIF) is a crucial task in embodied learning, requiring agents to interact with their environment through egocentric observations to fulfill natural language instructions. Recent advancements have seen a surge in employing large language models (LLMs) within a framework-centric approach to enhance performance in embodied learning tasks, including EIF. Despite these efforts, there exists a lack of a unified understanding regarding the impact of various components-ranging from visual perception to action execution-on task performance. To address this gap, we introduce OPEx, a comprehensive framework that delineates the core components essential for solving embodied learning tasks: Observer, Planner, and Executor. Through extensive evaluations, we provide a deep analysis of how each component influences EIF task performance. Furthermore, we innovate within this space by deploying a multi-agent dialogue strategy on a TextWorld counterpart, further enhancing task performance. Our findings reveal that LLM-centric design markedly improves EIF outcomes, identify visual perception and low-level action execution as critical bottlenecks, and demonstrate that augmenting LLMs with a multi-agent framework further elevates performance.

cs.AI↗

arXiv2024

Deciphering Digital Detectives: Understanding LLM Behaviors and Capabilities in Multi-Agent Mystery Games

In this study, we explore the application of Large Language Models (LLMs) in \textit{Jubensha}, a Chinese detective role-playing game and a novel area in Artificial Intelligence (AI) driven gaming. We introduce the first dataset specifically for Jubensha, including character scripts and game rules, to foster AI agent development in this complex narrative environment. Our work also presents a unique multi-agent interaction framework using LLMs, allowing AI agents to autonomously engage in this game. To evaluate the gaming performance of these AI agents, we developed novel methods measuring their mastery of case information and reasoning skills. Furthermore, we incorporated the latest advancements in in-context learning to improve the agents' performance in information gathering, murderer identification, and logical reasoning. The experimental results validate the effectiveness of our proposed methods. This work aims to offer a novel perspective on understanding LLM capabilities and establish a new benchmark for evaluating large language model-based agents.

cs.AI↗

arXiv2024

Bifurcation of time crystals in driven and dissipative Rydberg atomic gas

A time crystal is an exotic phase of matter where time-translational symmetry is broken; this phase differs from the spatial symmetry breaking induced in crystals in space. Lots of experiments report the transition from a thermal equilibrium phase to time crystal phase. However, there is no experimental method to probe the bifurcation effect of distinct time crystals in quantum many-body systems. Here, in a driven and dissipative many-body Rydberg atom system, we observe multiple continuous dissipative time crystals and emergence of more complex temporal symmetries beyond the single time crystal phase. Bifurcation of time crystals in strongly interacting Rydberg atoms is observed; the process manifests as a transition from a time crystal state of long temporal order to one of short temporal order, or vice versa. By manipulating the driving field parameters, we observe the time crystal's bistability and a hysteresis loop. These investigations indicate new possibilities for control and manipulation of the temporal symmetries of non-equilibrium systems.

cond-mat.quant-gas↗

arXiv2024

Microwave control of collective quantum jump statistics of a dissipative Rydberg gas

Quantum many-body systems near phase transitions respond collectively to externally applied perturbations. We explore this phenomenon in a laser-driven dissipative Rydberg gas that is tuned to a bistable regime. Here two metastable phases coexist, which feature a low and high density of Rydberg atoms, respectively. The ensuing collective dynamics, which we monitor in situ, is characterized by stochastic collective jumps between these two macroscopically distinct many-body phases. We show that the statistics of these jumps can be controlled using a dual-tone microwave field. In particular, we find that the distribution of jump times develops peaks corresponding to subharmonics of the relative microwave detuning. Our study demonstrates the control of collective statistical properties of dissipative quantum many-body systems without the necessity of fine-tuning or of ultra cold temperatures. Such robust phenomena may find technological applications in quantum sensing and metrology.

quant-ph↗

arXiv2024

GOAt: Explaining Graph Neural Networks via Graph Output Attribution

Understanding the decision-making process of Graph Neural Networks (GNNs) is crucial to their interpretability. Most existing methods for explaining GNNs typically rely on training auxiliary models, resulting in the explanations remain black-boxed. This paper introduces Graph Output Attribution (GOAt), a novel method to attribute graph outputs to input graph features, creating GNN explanations that are faithful, discriminative, as well as stable across similar samples. By expanding the GNN as a sum of scalar products involving node features, edge features and activation patterns, we propose an efficient analytical method to compute contribution of each node or edge feature to each scalar product and aggregate the contributions from all scalar products in the expansion form to derive the importance of each node and edge. Through extensive experiments on synthetic and real-world data, we show that our method not only outperforms various state-ofthe-art GNN explainers in terms of the commonly used fidelity metric, but also exhibits stronger discriminability, and stability by a remarkable margin.

cs.LG↗

arXiv2023

DenseShift: Towards Accurate and Efficient Low-Bit Power-of-Two Quantization

Efficiently deploying deep neural networks on low-resource edge devices is challenging due to their ever-increasing resource requirements. To address this issue, researchers have proposed multiplication-free neural networks, such as Power-of-Two quantization, or also known as Shift networks, which aim to reduce memory usage and simplify computation. However, existing low-bit Shift networks are not as accurate as their full-precision counterparts, typically suffering from limited weight range encoding schemes and quantization loss. In this paper, we propose the DenseShift network, which significantly improves the accuracy of Shift networks, achieving competitive performance to full-precision networks for vision and speech applications. In addition, we introduce a method to deploy an efficient DenseShift network using non-quantized floating-point activations, while obtaining 1.6X speed-up over existing methods. To achieve this, we demonstrate that zero-weight values in low-bit Shift networks do not contribute to model capacity and negatively impact inference computation. To address this issue, we propose a zero-free shifting mechanism that simplifies inference and increases model capacity. We further propose a sign-scale decomposition design to enhance training efficiency and a low-variance random initialization strategy to improve the model's transfer learning performance. Our extensive experiments on various computer vision and speech tasks demonstrate that DenseShift outperforms existing low-bit multiplication-free networks and achieves competitive performance compared to full-precision networks. Furthermore, our proposed approach exhibits strong transfer learning performance without a drop in accuracy. Our code was released on GitHub.

cs.CV↗

arXiv2023

HoneyBee: Progressive Instruction Finetuning of Large Language Models for Materials Science

We propose an instruction-based process for trustworthy data curation in materials science (MatSci-Instruct), which we then apply to finetune a LLaMa-based language model targeted for materials science (HoneyBee). MatSci-Instruct helps alleviate the scarcity of relevant, high-quality materials science textual data available in the open literature, and HoneyBee is the first billion-parameter language model specialized to materials science. In MatSci-Instruct we improve the trustworthiness of generated data by prompting multiple commercially available large language models for generation with an Instructor module (e.g. Chat-GPT) and verification from an independent Verifier module (e.g. Claude). Using MatSci-Instruct, we construct a dataset of multiple tasks and measure the quality of our dataset along multiple dimensions, including accuracy against known facts, relevance to materials science, as well as completeness and reasonableness of the data. Moreover, we iteratively generate more targeted instructions and instruction-data in a finetuning-evaluation-feedback loop leading to progressively better performance for our finetuned HoneyBee models. Our evaluation on the MatSci-NLP benchmark shows HoneyBee's outperformance of existing language models on materials science tasks and iterative improvement in successive stages of instruction-data refinement. We study the quality of HoneyBee's language modeling through automatic evaluation and analyze case studies to further understand the model's capabilities and limitations. Our code and relevant datasets are publicly available at \url{https://github.com/BangLab-UdeM-Mila/NLP4MatSci-HoneyBee}.

cs.CL↗

arXiv2023

Multimodal Multi-Hop Question Answering Through a Conversation Between Tools and Efficiently Finetuned Large Language Models

We employ a tool-interacting divide-and-conquer strategy enabling large language models (LLMs) to answer complex multimodal multi-hop questions. In particular, we harness the power of large language models to divide a given multimodal multi-hop question into unimodal single-hop sub-questions to be answered by the appropriate tool from a predefined set of tools. After all corresponding tools provide the LLM with their answers, the LLM generates the next relevant unimodal single-hop question. To increase the reasoning ability of LLMs, we prompt chatGPT to generate a tool-interacting divide-and-conquer dataset. This dataset is then used to efficiently finetune the corresponding LLM. To assess the effectiveness of this approach, we conduct an evaluation on two recently introduced complex question-answering datasets. The experimental analysis demonstrate substantial improvements over existing state-of-the-art solutions, indicating the efficacy and generality of our strategy

cs.CL↗

arXiv2023

Progress on two-dimensional ferrovalley materials

The electron's charge and spin degrees of freedom are at the core of modern electronic devices. With the in-depth investigation of two-dimensional materials, another degree of freedom, valley, has also attracted tremendous research interest. The intrinsic spontaneous valley polarization in two-dimensional magnetic systems, ferrovalley material, provides convenience for detecting and modulating the valley. In this review, we first introduce the development of valleytronics. Then, the valley polarization forms by the p, d, and f-orbit that are discussed. Following, we discuss the investigation progress of modulating the valley polarization of two-dimensional ferrovalley materials by multiple physical fields, such as electric, stacking mode, strain, and interface. Finally, we look forward to the future developments of valleytronics.

cond-mat.mtrl-sci↗

arXiv2023

Fine-tuning Happens in Tiny Subspaces: Exploring Intrinsic Task-specific Subspaces of Pre-trained Language Models

Pre-trained language models (PLMs) are known to be overly parameterized and have significant redundancy, indicating a small degree of freedom of the PLMs. Motivated by the observation, in this paper, we study the problem of re-parameterizing and fine-tuning PLMs from a new perspective: Discovery of intrinsic task-specific subspace. Specifically, by exploiting the dynamics of the fine-tuning process for a given task, the parameter optimization trajectory is learned to uncover its intrinsic task-specific subspace. A key finding is that PLMs can be effectively fine-tuned in the subspace with a small number of free parameters. Beyond, we observe some outlier dimensions emerging during fine-tuning in the subspace. Disabling these dimensions degrades the model performance significantly. This suggests that these dimensions are crucial to induce task-specific knowledge to downstream tasks.

cs.CL↗

arXiv2023

Ni-O-Ag catalyst enables 103-m$^2$ artificial photosynthesis with >16% solar-to-chemical energy conversion efficiency

Herein, NiO nanosheets supported with Ag single atoms are synthesized for photothermal CO2 hydrogenation to achieve 1065 mmol g$^{-1}$ h$^{-1}$ of CO production rate under 1 sun irradiation, revealing the unparalleled weak sunlight driven reverse water-gas shift reaction (RWGS) activity. This performance is attributed to the coupling effect of Ag-O-Ni sites to enhance the hydrogenation of CO$_2$ and weaken the CO adsorption, resulting in 1434 mmol g$^{-1}$ h$^{-1}$ of CO yield at 300$^\circ$ C, surpassing any low-temperature RWGS performances ever reported. Building on this, we integrated the 2D Ni$_1$Ag$_{0.02}$O$_1$ supported photothermal RWGS with commercial photovoltaic electrolytic water splitting, leading to the realization of 103 m$^2$ scale artificial photosynthesis system (CO$_2$+H$_2$$\to$CO+H$_2$O) with a daily CO yield of 18.70 m$^3$, a photochemical energy conversion efficiency of >16%, over 90% H$_2$ ultilization efficiency, outperforming other types of artificial photosynthesis. The results of this research chart a promising course for designing practical, natural sunlight-driven artificial photosynthesis systems and highly efficient platinum-free CO$_2$ hydrogenation catalysts. This work is a significant step towards harnessing solar energy more efficiently and sustainably, opening exciting possibilities for future research and development in this area.

cond-mat.mtrl-sci↗

arXiv2023

Electric field measurement and application based on Rydberg atoms

Microwave sensing has important applications in areas such as data communication and remote sensing, so it has received much attention from international academia, industry, and governments. Atomic wireless sensing uses the strong response of the large electric dipole moment of a Rydberg atom to an external field to achieve precise measurement of a radio frequency (RF) electric field. This has advantages over traditional wireless sensing. The advantage of a Rydberg atom is its ultra-wide energy level transitions, which make it responsive to RF electric fields over a wide bandwidth. Here, we briefly review the progress of electric field measurement based on Rydberg atoms. The main contents include the properties of Rydberg atoms, measurement using Rydberg atoms, and experimental progress in electric field measurement in different bands. We show the different methods for detecting electric fields such as atomic superheterodyne, machine learning, and critically enhanced measurement. The development of Rydberg atomic measurement focuses on the advantages of Rydberg atomic sensing, especially compared with conventional microwave receivers. This is of major significance to developing Rydberg-based measurement in astronomy, remote sensing, and other fields.

physics.atom-ph↗

arXiv2023

MatSci-NLP: Evaluating Scientific Language Models on Materials Science Language Tasks Using Text-to-Schema Modeling

We present MatSci-NLP, a natural language benchmark for evaluating the performance of natural language processing (NLP) models on materials science text. We construct the benchmark from publicly available materials science text data to encompass seven different NLP tasks, including conventional NLP tasks like named entity recognition and relation classification, as well as NLP tasks specific to materials science, such as synthesis action retrieval which relates to creating synthesis procedures for materials. We study various BERT-based models pretrained on different scientific text corpora on MatSci-NLP to understand the impact of pretraining strategies on understanding materials science text. Given the scarcity of high-quality annotated data in the materials science domain, we perform our fine-tuning experiments with limited training data to encourage the generalize across MatSci-NLP tasks. Our experiments in this low-resource training setting show that language models pretrained on scientific text outperform BERT trained on general text. MatBERT, a model pretrained specifically on materials science journals, generally performs best for most tasks. Moreover, we propose a unified text-to-schema for multitask learning on \benchmark and compare its performance with traditional fine-tuning methods. In our analysis of different training methods, we find that our proposed text-to-schema methods inspired by question-answering consistently outperform single and multitask NLP fine-tuning methods. The code and datasets are publicly available at \url{https://github.com/BangLab-UdeM-Mila/NLP4MatSci-ACL23}.

cs.CL↗

arXiv2023

SkillQG: Learning to Generate Question for Reading Comprehension Assessment

We present $\textbf{$\texttt{SkillQG}$}$: a question generation framework with controllable comprehension types for assessing and improving machine reading comprehension models. Existing question generation systems widely differentiate questions by $\textit{literal}$ information such as question words and answer types to generate semantically relevant questions for a given context. However, they rarely consider the $\textit{comprehension}$ nature of questions, i.e. the different comprehension capabilities embodied by different questions. In comparison, our $\texttt{SkillQG}$ is able to tailor a fine-grained assessment and improvement to the capabilities of question answering models built on it. Specifically, we first frame the comprehension type of questions based on a hierarchical skill-based schema, then formulate $\texttt{SkillQG}$ as a skill-conditioned question generator. Furthermore, to improve the controllability of generation, we augment the input text with question focus and skill-specific knowledge, which are constructed by iteratively prompting the pre-trained language models. Empirical results demonstrate that $\texttt{SkillQG}$ outperforms baselines in terms of quality, relevance, and skill-controllability while showing a promising performance boost in downstream question answering task.

cs.CL↗

arXiv2022

MCSCSet: A Specialist-annotated Dataset for Medical-domain Chinese Spelling Correction

Chinese Spelling Correction (CSC) is gaining increasing attention due to its promise of automatically detecting and correcting spelling errors in Chinese texts. Despite its extensive use in many applications, like search engines and optical character recognition systems, little has been explored in medical scenarios in which complex and uncommon medical entities are easily misspelled. Correcting the misspellings of medical entities is arguably more difficult than those in the open domain due to its requirements of specificdomain knowledge. In this work, we define the task of Medical-domain Chinese Spelling Correction and propose MCSCSet, a large scale specialist-annotated dataset that contains about 200k samples. In contrast to the existing open-domain CSC datasets, MCSCSet involves: i) extensive real-world medical queries collected from Tencent Yidian, ii) corresponding misspelled sentences manually annotated by medical specialists. To ensure automated dataset curation, MCSCSet further offers a medical confusion set consisting of the commonly misspelled characters of given Chinese medical terms. This enables one to create the medical misspelling dataset automatically. Extensive empirical studies have shown significant performance gaps between the open-domain and medical-domain spelling correction, highlighting the need to develop high-quality datasets that allow for Chinese spelling correction in specific domains. Moreover, our work benchmarks several representative Chinese spelling correction models, establishing baselines for future work.

cs.CL↗

arXiv2022

Improving Multi-turn Emotional Support Dialogue Generation with Lookahead Strategy Planning

Providing Emotional Support (ES) to soothe people in emotional distress is an essential capability in social interactions. Most existing researches on building ES conversation systems only considered single-turn interactions with users, which was over-simplified. In comparison, multi-turn ES conversation systems can provide ES more effectively, but face several new technical challenges, including: (1) how to adopt appropriate support strategies to achieve the long-term dialogue goal of comforting the user's emotion; (2) how to dynamically model the user's state. In this paper, we propose a novel system MultiESC to address these issues. For strategy planning, drawing inspiration from the A* search algorithm, we propose lookahead heuristics to estimate the future user feedback after using particular strategies, which helps to select strategies that can lead to the best long-term effects. For user state modeling, MultiESC focuses on capturing users' subtle emotional expressions and understanding their emotion causes. Extensive experiments show that MultiESC significantly outperforms competitive baselines in both dialogue generation and strategy planning. Our codes are available at https://github.com/lwgkzl/MultiESC.

cs.CL↗

arXiv2022

Attend and select: A segment selective transformer for microblog hashtag generation

Hashtag generation aims to generate short and informal topical tags from a microblog post, in which tokens or phrases form the hashtags. These tokens or phrases may originate from primary fragmental textual pieces (e.g., segments) in the original text and are separated into different segments. However, conventional sequence-to-sequence generation methods are hard to filter out secondary information from different textual granularity and are not good at selecting crucial tokens. Thus, they are suboptimal in generating more condensed hashtags. In this work, we propose a modified Transformer-based generation model with adding a segments-selection procedure for the original encoding and decoding phases. The segments-selection phase is based on a novel Segments Selection Mechanism (SSM) to model different textual granularity on global text, local segments, and tokens, contributing to generating condensed hashtags. Specifically, it first attends to primary semantic segments and then transforms discontinuous segments from the source text into a sequence of hashtags by selecting crucial tokens. Extensive evaluations on the two datasets reveal our approach's superiority with significant improvements to the extraction and generation baselines. The code and datasets are available at https://github.com/OpenSUM/HashtagGen.

cs.CL↗

arXiv2022

Highly sensitive measurement of a megahertz rf electric field with a Rydberg-atom sensor

Rydberg atoms have great potential in electric field measurement and have an advantage with a large frequency bandwidth from the kHz to the THz scale. However, the sensitivity for measuring a weak MHz electric field signal is limited by the spectroscopic resolution, because the weak electric field induces only a small perturbation of the population and energy level shift of the Rydberg atoms. Here, we report highly sensitive measurement of a weak MHz electric field using electromagnetically induced transparency with Rydberg atoms in a thermal atomic system. Using the heterodyne method on a 30-MHz electric field, we successfully measure the minimum electric field strength to be \textcolor{black}{37.3 $\mathrm{μV/cm}$} with a sensitivity up to $-65$ dBm/Hz and a linear dynamic range over 65 dB. Additionally, we measure an amplitude-modulated signal and demodulate the signal with a fidelity over 98\%. This work extends the sensitivity of atomic sensors for measuring MHz electric fields, which advances atomic electric field-sensing technology.

physics.atom-ph↗