arXiv subjects

Pengfei Liu

Publications and source records attributed to Pengfei Liu.

At least 181 records · Page 10Linked to original sources

arXiv2023

Align on the Fly: Adapting Chatbot Behavior to Established Norms

In this paper, we aim to align large language models with the ever-changing, complex, and diverse human values (e.g., social norms) across time and locations. This presents a challenge to existing alignment techniques, such as supervised fine-tuning, which internalize values within model parameters. To overcome this, we propose an On-the-fly Preference Optimization (OPO) method, which is a real-time alignment that works in a streaming way. It employs an external memory to store established rules for alignment, which can constrain LLMs' behaviors without further training, allowing for convenient updates and customization of human values. We also introduce a scalable evaluation to assess the proposed method more effectively. Experimental results on both human-annotated and auto-generated questions from legal and moral domains indicate the effectiveness of the proposed OPO method. Our code and data are released at https://github.com/GAIR-NLP/OPO.

cs.CL↗

arXiv2023

Chiral Dirac fermion in a collinear antiferromagnet

In a Dirac semimetal, the massless Dirac fermion has zero chirality, leading to surface states connected adiabatically to a topologically trivial surface state as well as vanishing anomalous Hall effect (AHE). Recently, it is predicted that in the nonrelativistic limit of certain collinear antiferromagnets, there exists a type of chiral Dirac-like fermion, whose dispersion manifests four-fold degenerate crossing points formed by spin-degenerate linear bands, with topologically protected Fermi arcs. Such unconventional chiral fermion, protected by a hidden SU(2) symmetry in the hierarchy of an enhanced crystallographic group, namely spin space group, is not experimentally verified yet. Here, by angle-resolved photoemission spectroscopy measurements, we reveal the surface origin of the electron pocket at the Fermi surface in collinear antiferromagnet CoNb3S6. Combining with neutron diffraction and first-principles calculations, we suggest a multidomain collinear AFM configuration, rendering the the existence of the Fermi-arc surface states induced by chiral Dirac-like fermions. Our work provides spectral evidence of the chiral Dirac-like fermion caused by particular spin symmetry in CoNb3S6, paving an avenue for exploring new emergent phenomena in antiferromagnets with unconventional quasiparticle excitations.

cond-mat.mtrl-sci↗

arXiv2023

Generative Judge for Evaluating Alignment

The rapid development of Large Language Models (LLMs) has substantially expanded the range of tasks they can address. In the field of Natural Language Processing (NLP), researchers have shifted their focus from conventional NLP tasks (e.g., sequence tagging and parsing) towards tasks that revolve around aligning with human needs (e.g., brainstorming and email writing). This shift in task distribution imposes new requirements on evaluating these aligned models regarding generality (i.e., assessing performance across diverse scenarios), flexibility (i.e., examining under different protocols), and interpretability (i.e., scrutinizing models with explanations). In this paper, we propose a generative judge with 13B parameters, Auto-J, designed to address these challenges. Our model is trained on user queries and LLM-generated responses under massive real-world scenarios and accommodates diverse evaluation protocols (e.g., pairwise response comparison and single-response evaluation) with well-structured natural language critiques. To demonstrate the efficacy of our approach, we construct a new testbed covering 58 different scenarios. Experimentally, Auto-J outperforms a series of strong competitors, including both open-source and closed-source models, by a large margin. We also provide detailed analysis and case studies to further reveal the potential of our method and make a variety of resources public at https://github.com/GAIR-NLP/auto-j.

cs.CL↗

arXiv2023

FELM: Benchmarking Factuality Evaluation of Large Language Models

Assessing factuality of text generated by large language models (LLMs) is an emerging yet crucial research area, aimed at alerting users to potential errors and guiding the development of more reliable LLMs. Nonetheless, the evaluators assessing factuality necessitate suitable evaluation themselves to gauge progress and foster advancements. This direction remains under-explored, resulting in substantial impediments to the progress of factuality evaluators. To mitigate this issue, we introduce a benchmark for Factuality Evaluation of large Language Models, referred to as felm. In this benchmark, we collect responses generated from LLMs and annotate factuality labels in a fine-grained manner. Contrary to previous studies that primarily concentrate on the factuality of world knowledge (e.g.~information from Wikipedia), felm focuses on factuality across diverse domains, spanning from world knowledge to math and reasoning. Our annotation is based on text segments, which can help pinpoint specific factual errors. The factuality annotations are further supplemented by predefined error types and reference links that either support or contradict the statement. In our experiments, we investigate the performance of several LLM-based factuality evaluators on felm, including both vanilla LLMs and those augmented with retrieval mechanisms and chain-of-thought processes. Our findings reveal that while retrieval aids factuality evaluation, current LLMs are far from satisfactory to faithfully detect factual errors.

cs.CL↗

arXiv2023

Towards Interpretable and Efficient Automatic Reference-Based Summarization Evaluation

Interpretability and efficiency are two important considerations for the adoption of neural automatic metrics. In this work, we develop strong-performing automatic metrics for reference-based summarization evaluation, based on a two-stage evaluation pipeline that first extracts basic information units from one text sequence and then checks the extracted units in another sequence. The metrics we developed include two-stage metrics that can provide high interpretability at both the fine-grained unit level and summary level, and one-stage metrics that achieve a balance between efficiency and interpretability. We make the developed tools publicly available at https://github.com/Yale-LILY/AutoACU.

cs.CL↗

arXiv2023

Let's reward step by step: Step-Level reward model as the Navigators for Reasoning

Recent years have seen considerable advancements in multi-step reasoning with Large Language Models (LLMs). The previous studies have elucidated the merits of integrating feedback or search mechanisms during model inference to improve the reasoning accuracy. The Process-Supervised Reward Model (PRM), typically furnishes LLMs with step-by-step feedback during the training phase, akin to Proximal Policy Optimization (PPO) or reject sampling. Our objective is to examine the efficacy of PRM in the inference phase to help discern the optimal solution paths for multi-step tasks such as mathematical reasoning and code generation. To this end, we propose a heuristic greedy search algorithm that employs the step-level feedback from PRM to optimize the reasoning pathways explored by LLMs. This tailored PRM demonstrated enhanced results compared to the Chain of Thought (CoT) on mathematical benchmarks like GSM8K and MATH. Additionally, to explore the versatility of our approach, we develop a novel method to automatically generate step-level reward dataset for coding tasks and observed similar improved performance in the code generation tasks. Thus highlighting the robust nature of our reward-model-based approach to inference for reasoning tasks.

cs.CL↗

arXiv2023

Helically-trapped electron mode in optimized stellarator

Global gyrokinetic simulations find a strong helically-trapped electron mode (HTEM) driven by density gradients in the optimized stellarator W7-X fusion experiment. The eigenmode structure localizes in the inner side of the torus with an unfavorable magnetic curvature and weak magnetic field, where there is a large fraction of helically-trapped electrons. The mode is excited mostly by the ion free energy and propagates in the ion direction with a linear frequency much smaller than the diamagnetic frequency. The instability saturates by nonlinear processes of turbulence spreading in the real space and spectral transfer from unstable to damped regions. The steady state HTEM turbulence drives a large particle flux that may have significant implications for the confinement of fusion fuel and removal of fusion ash in the stellarator reactor.

physics.plasm-ph↗

arXiv2023

Being Aware of Localization Accuracy By Generating Predicted-IoU-Guided Quality Scores

Localization Quality Estimation (LQE) helps to improve detection performance as it benefits post processing through jointly considering classification score and localization accuracy. In this perspective, for further leveraging the close relationship between localization accuracy and IoU (Intersection-Over-Union), and for depressing those inconsistent predictions, we designed an elegant LQE branch to acquire localization quality score guided by predicted IoU. Distinctly, for alleviating the inconsistency of classification score and localization quality during training and inference, under which some predictions with low classification scores but high LQE scores will impair the performance, instead of separately and independently setting, we embedded LQE branch into classification branch, producing a joint classification-localization-quality representation. Then a novel one stage detector termed CLQ is proposed. Extensive experiments show that CLQ achieves state-of-the-arts' performance at an accuracy of 47.8 AP and a speed of 11.5 fps with ResNeXt-101 as backbone on COCO test-dev. Finally, we extend CLQ to ATSS, producing a reliable 1.2 AP gain, showing our model's strong adaptability and scalability. Codes are released at https://github.com/PanffeeReal/CLQ.

cs.CV↗

arXiv2023

The Impact of Downgrading Protected Areas (PAD) on Biodiversity

We quantitatively assess the impacts of Downgrading Protected Areas (PAD) on biodiversity in the U.S.. Results show that PAD events significantly reduce biodiversity. The proximity to PAD events decreases the biodiversity by 26.0% within 50 km compared with records of species further away from the PAD events. We observe an overall 32.3% decrease in abundance after those nearest PAD events are enacted. Abundance declines more in organisms living in contact with water and non-mammals. Species abundance is more sensitive to the negative impacts in areas where PAD events were later reversed, as well as in areas close to protected areas belonging to the International Union for Conservation of Nature (IUCN) category. The enacted PAD events between the period 1903 to 2018 in the U.S. lead to economic losses of approximately $689.95 million due to decrease in abundance. Our results contribute to the understanding on the impact of environmental interventions such as PAD events on biodiversity change and provide important implications on biodiversity conservation policies.

q-bio.PE↗

arXiv2023

FacTool: Factuality Detection in Generative AI -- A Tool Augmented Framework for Multi-Task and Multi-Domain Scenarios

The emergence of generative pre-trained models has facilitated the synthesis of high-quality text, but it has also posed challenges in identifying factual errors in the generated text. In particular: (1) A wider range of tasks now face an increasing risk of containing factual errors when handled by generative models. (2) Generated texts tend to be lengthy and lack a clearly defined granularity for individual facts. (3) There is a scarcity of explicit evidence available during the process of fact checking. With the above challenges in mind, in this paper, we propose FacTool, a task and domain agnostic framework for detecting factual errors of texts generated by large language models (e.g., ChatGPT). Experiments on four different tasks (knowledge-based QA, code generation, mathematical reasoning, and scientific literature review) show the efficacy of the proposed method. We release the code of FacTool associated with ChatGPT plugin interface at https://github.com/GAIR-NLP/factool .

cs.CL↗

arXiv2023

Improving Factuality of Abstractive Summarization via Contrastive Reward Learning

Modern abstractive summarization models often generate summaries that contain hallucinated or contradictory information. In this paper, we propose a simple but effective contrastive learning framework that incorporates recent developments in reward learning and factuality metrics. Empirical studies demonstrate that the proposed framework enables summarization models to learn from feedback of factuality metrics using contrastive reward learning, leading to more factual summaries by human evaluations. This suggests that further advances in learning and evaluation algorithms can feed directly into providing more factual summaries.

cs.CL↗

arXiv2023

DataFinder: Scientific Dataset Recommendation from Natural Language Descriptions

Modern machine learning relies on datasets to develop and validate research ideas. Given the growth of publicly available data, finding the right dataset to use is increasingly difficult. Any research question imposes explicit and implicit constraints on how well a given dataset will enable researchers to answer this question, such as dataset size, modality, and domain. We operationalize the task of recommending datasets given a short natural language description of a research idea, to help people find relevant datasets for their needs. Dataset recommendation poses unique challenges as an information retrieval problem; datasets are hard to directly index for search and there are no corpora readily available for this task. To facilitate this task, we build the DataFinder Dataset which consists of a larger automatically-constructed training set (17.5K queries) and a smaller expert-annotated evaluation set (392 queries). Using this data, we compare various information retrieval algorithms on our test set and present a superior bi-encoder retriever for text-based dataset recommendation. This system, trained on the DataFinder Dataset, finds more relevant search results than existing third-party dataset search engines. To encourage progress on dataset recommendation, we release our dataset and models to the public.

cs.IR↗

arXiv2023

Revisiting the Gold Standard: Grounding Summarization Evaluation with Robust Human Evaluation

Human evaluation is the foundation upon which the evaluation of both summarization systems and automatic metrics rests. However, existing human evaluation studies for summarization either exhibit a low inter-annotator agreement or have insufficient scale, and an in-depth analysis of human evaluation is lacking. Therefore, we address the shortcomings of existing summarization evaluation along the following axes: (1) We propose a modified summarization salience protocol, Atomic Content Units (ACUs), which is based on fine-grained semantic units and allows for a high inter-annotator agreement. (2) We curate the Robust Summarization Evaluation (RoSE) benchmark, a large human evaluation dataset consisting of 22,000 summary-level annotations over 28 top-performing systems on three datasets. (3) We conduct a comparative study of four human evaluation protocols, underscoring potential confounding factors in evaluation setups. (4) We evaluate 50 automatic metrics and their variants using the collected human annotations across evaluation protocols and demonstrate how our benchmark leads to more statistically stable and significant results. The metrics we benchmarked include recent methods based on large language models (LLMs), GPTScore and G-Eval. Furthermore, our findings have important implications for evaluating LLMs, as we show that LLMs adjusted by human feedback (e.g., GPT-3.5) may overfit unconstrained human evaluation, which is affected by the annotators' prior, input-agnostic preferences, calling for more robust, targeted evaluation methods.

cs.CL↗

arXiv2023

Multi-Dimensional Evaluation of Text Summarization with In-Context Learning

Evaluation of natural language generation (NLG) is complex and multi-dimensional. Generated text can be evaluated for fluency, coherence, factuality, or any other dimensions of interest. Most frameworks that perform such multi-dimensional evaluation require training on large manually or synthetically generated datasets. In this paper, we study the efficacy of large language models as multi-dimensional evaluators using in-context learning, obviating the need for large training datasets. Our experiments show that in-context learning-based evaluators are competitive with learned evaluation frameworks for the task of text summarization, establishing state-of-the-art on dimensions such as relevance and factual consistency. We then analyze the effects of factors such as the selection and number of in-context examples on performance. Finally, we study the efficacy of in-context learning based evaluators in evaluating zero-shot summaries written by large language models such as GPT-3.

cs.CL↗

arXiv2023

GlobalBench: A Benchmark for Global Progress in Natural Language Processing

Despite the major advances in NLP, significant disparities in NLP system performance across languages still exist. Arguably, these are due to uneven resource allocation and sub-optimal incentives to work on less resourced languages. To track and further incentivize the global development of equitable language technology, we introduce GlobalBench. Prior multilingual benchmarks are static and have focused on a limited number of tasks and languages. In contrast, GlobalBench is an ever-expanding collection that aims to dynamically track progress on all NLP datasets in all languages. Rather than solely measuring accuracy, GlobalBench also tracks the estimated per-speaker utility and equity of technology across all languages, providing a multi-faceted view of how language technology is serving people of the world. Furthermore, GlobalBench is designed to identify the most under-served languages, and rewards research efforts directed towards those languages. At present, the most under-served languages are the ones with a relatively high population, but nonetheless overlooked by composite multilingual benchmarks (like Punjabi, Portuguese, and Wu Chinese). Currently, GlobalBench covers 966 datasets in 190 languages, and has 1,128 system submissions spanning 62 languages.

cs.CL↗

arXiv2023

LIMA: Less Is More for Alignment

Large language models are trained in two stages: (1) unsupervised pretraining from raw text, to learn general-purpose representations, and (2) large scale instruction tuning and reinforcement learning, to better align to end tasks and user preferences. We measure the relative importance of these two stages by training LIMA, a 65B parameter LLaMa language model fine-tuned with the standard supervised loss on only 1,000 carefully curated prompts and responses, without any reinforcement learning or human preference modeling. LIMA demonstrates remarkably strong performance, learning to follow specific response formats from only a handful of examples in the training data, including complex queries that range from planning trip itineraries to speculating about alternate history. Moreover, the model tends to generalize well to unseen tasks that did not appear in the training data. In a controlled human study, responses from LIMA are either equivalent or strictly preferred to GPT-4 in 43% of cases; this statistic is as high as 58% when compared to Bard and 65% versus DaVinci003, which was trained with human feedback. Taken together, these results strongly suggest that almost all knowledge in large language models is learned during pretraining, and only limited instruction tuning data is necessary to teach models to produce high quality output.

cs.CL↗

arXiv2023

MusicFace: Music-driven Expressive Singing Face Synthesis

It is still an interesting and challenging problem to synthesize a vivid and realistic singing face driven by music signal. In this paper, we present a method for this task with natural motions of the lip, facial expression, head pose, and eye states. Due to the coupling of the mixed information of human voice and background music in common signals of music audio, we design a decouple-and-fuse strategy to tackle the challenge. We first decompose the input music audio into human voice stream and background music stream. Due to the implicit and complicated correlation between the two-stream input signals and the dynamics of the facial expressions, head motions and eye states, we model their relationship with an attention scheme, where the effects of the two streams are fused seamlessly. Furthermore, to improve the expressiveness of the generated results, we propose to decompose head movements generation into speed generation and direction generation, and decompose eye states generation into the short-time eye blinking generation and the long-time eye closing generation to model them separately. We also build a novel SingingFace Dataset to support the training and evaluation of this task, and to facilitate future works on this topic. Extensive experiments and user study show that our proposed method is capable of synthesizing vivid singing face, which is better than state-of-the-art methods qualitatively and quantitatively.

cs.GR↗

arXiv2023

GPTScore: Evaluate as You Desire

Generative Artificial Intelligence (AI) has enabled the development of sophisticated models that are capable of producing high-caliber text, images, and other outputs through the utilization of large pre-trained models. Nevertheless, assessing the quality of the generation is an even more arduous task than the generation itself, and this issue has not been given adequate consideration recently. This paper proposes a novel evaluation framework, GPTScore, which utilizes the emergent abilities (e.g., zero-shot instruction) of generative pre-trained models to score generated texts. There are 19 pre-trained models explored in this paper, ranging in size from 80M (e.g., FLAN-T5-small) to 175B (e.g., GPT3). Experimental results on four text generation tasks, 22 evaluation aspects, and corresponding 37 datasets demonstrate that this approach can effectively allow us to achieve what one desires to evaluate for texts simply by natural language instructions. This nature helps us overcome several long-standing challenges in text evaluation--how to achieve customized, multi-faceted evaluation without the need for annotated samples. We make our code publicly available at https://github.com/jinlanfu/GPTScore.

cs.CL↗