Searcharxiv⌕ Search

arXiv subjects

Xiao Liu

Publications and source records attributed to Xiao Liu.

At least 325 records · Page 18Linked to original sources
arXiv2023

Highly Accurate Prediction of NMR Chemical Shifts from Low-Level Quantum Mechanics Calculations Using Machine Learning

Theoretical predictions of NMR chemical shifts from first-principles can greatly facilitate experimental interpretation and structure identification. However, accurate prediction of chemical shifts using the best coupled cluster methods can be prohibitively expensive for systems larger than ten to twenty non-hydrogen atoms on today's computers. By contrast machine learning methods offer inexpensive alternatives but are hampered by generalization to molecules outside the original training set. Here we propose a novel machine learning feature representation informed by intermediate calculations of atomic chemical shielding tensors within a molecular environment using an inexpensive quantum mechanics method, and training it to predict NMR chemical shieldings of a high-level composite theory that is comparable to CCSD(T) in the complete basis set limit. The inexpensive shift machine learning (iShiftML) algorithm is trained through a new progressive active learning workflow that reduces the total number of expensive calculations required when constructing the dataset, while allowing the model to continuously improve on data it has never seen. Furthermore, we show that the error estimations from our model correlate quite well with actual errors to provide confidence values on new predictions. We illustrate the predictive capacity of iShiftML across gas phase experimental chemical shifts for small organic molecules and much larger and more complex natural products in which we can accurately differentiate between subtle diastereomers based on chemical shift assignments.

physics.chem-ph↗
arXiv2023

WebGLM: Towards An Efficient Web-Enhanced Question Answering System with Human Preferences

We present WebGLM, a web-enhanced question-answering system based on the General Language Model (GLM). Its goal is to augment a pre-trained large language model (LLM) with web search and retrieval capabilities while being efficient for real-world deployments. To achieve this, we develop WebGLM with strategies for the LLM-augmented retriever, bootstrapped generator, and human preference-aware scorer. Specifically, we identify and address the limitations of WebGPT (OpenAI), through which WebGLM is enabled with accuracy, efficiency, and cost-effectiveness advantages. In addition, we propose systematic criteria for evaluating web-enhanced QA systems. We conduct multi-dimensional human evaluation and quantitative ablation studies, which suggest the outperformance of the proposed WebGLM designs over existing systems. WebGLM with the 10-billion-parameter GLM (10B) is shown to perform better than the similar-sized WebGPT (13B) and even comparably to WebGPT (175B) in human evaluation. The code, demo, and data are at \url{https://github.com/THUDM/WebGLM}.

cs.CL↗
arXiv2023

Does Debiasing Inevitably Degrade the Model Performance

Gender bias in language models has attracted sufficient attention because it threatens social justice. However, most of the current debiasing methods degraded the model's performance on other tasks while the degradation mechanism is still mysterious. We propose a theoretical framework explaining the three candidate mechanisms of the language model's gender bias. We use our theoretical framework to explain why the current debiasing methods cause performance degradation. We also discover a pathway through which debiasing will not degrade the model performance. We further develop a causality-detection fine-tuning approach to correct gender bias. The numerical experiment demonstrates that our method is able to lead to double dividends: partially mitigating gender bias while avoiding performance degradation.

cs.CL↗
arXiv2023

Multi-Modal Emotion Recognition for Enhanced Requirements Engineering: A Novel Approach

Requirements engineering (RE) plays a crucial role in developing software systems by bridging the gap between stakeholders' needs and system specifications. However, effective communication and elicitation of stakeholder requirements can be challenging, as traditional RE methods often overlook emotional cues. This paper introduces a multi-modal emotion recognition platform (MEmoRE) to enhance the requirements engineering process by capturing and analyzing the emotional cues of stakeholders in real-time. MEmoRE leverages state-of-the-art emotion recognition techniques, integrating facial expression, vocal intonation, and textual sentiment analysis to comprehensively understand stakeholder emotions. This multi-modal approach ensures the accurate and timely detection of emotional cues, enabling requirements engineers to tailor their elicitation strategies and improve overall communication with stakeholders. We further intend to employ our platform for later RE stages, such as requirements reviews and usability testing. By integrating multi-modal emotion recognition into requirements engineering, we aim to pave the way for more empathetic, effective, and successful software development processes. We performed a preliminary evaluation of our platform. This paper reports on the platform design, preliminary evaluation, and future development plan as an ongoing project.

cs.SE↗
arXiv2023

How Many Answers Should I Give? An Empirical Study of Multi-Answer Reading Comprehension

The multi-answer phenomenon, where a question may have multiple answers scattered in the document, can be well handled by humans but is challenging enough for machine reading comprehension (MRC) systems. Despite recent progress in multi-answer MRC, there lacks a systematic analysis of how this phenomenon arises and how to better address it. In this work, we design a taxonomy to categorize commonly-seen multi-answer MRC instances, with which we inspect three multi-answer datasets and analyze where the multi-answer challenge comes from. We further analyze how well different paradigms of current multi-answer MRC models deal with different types of multi-answer instances. We find that some paradigms capture well the key information in the questions while others better model the relationship between questions and contexts. We thus explore strategies to make the best of the strengths of different paradigms. Experiments show that generation models can be a promising platform to incorporate different paradigms. Our annotations and code are released for further research.

cs.CL↗
arXiv2023

On the spectra of the gravity water waves linearized at monotone shear flows

We consider the spectra of the 2-dim gravity waves of finite depth linearized at a uniform monotonic shear flow $U(x_2)$, $x_2 \in (-h, 0)$, where the wave numbers $k$ of the horizontal variable $x_1$ is treated as a parameter. Our main results include a.) a complete branch of non-singular neutral modes $c^+(k)$ strictly decreasing in $k\ge 0$ and converging to $U(0)$ as $k \to \infty$; b.) another branch of non-singular neutral modes $c_-(k)$, $k \in (-k_-, k_-)$ for some $k_->0$, with $c_-(\pm k_-) = U(-h)$; c.) the non-degeneracy and the bifurcation at $(k_-, c=U(-h))$; d.) the existence and non-existence of unstable modes for $c$ near $U(0)$, $U(-h)$, and interior inflection values of $U$; e.) the complete spectral distribution in the case where $U''$ does not change sign or changes sign exactly once non-degenerately. In particular, $U$ is spectrally stable if $U'U''\le 0$ and unstable if $U$ has a non-degenerate interior inflection value or $\{U'U''>0\}$ accumulate at $x_2=-h$ or $0$. Moreover, if $U$ is an unstable shear flow of the fixed boundary problem in a channel, then strong gravity could cause instability of the linearized gravity waves in all long waves (i.e. $|k|\ll1$).

math.AP↗
arXiv2023

Contrastive Hierarchical Discourse Graph for Scientific Document Summarization

The extended structural context has made scientific paper summarization a challenging task. This paper proposes CHANGES, a contrastive hierarchical graph neural network for extractive scientific paper summarization. CHANGES represents a scientific paper with a hierarchical discourse graph and learns effective sentence representations with dedicated designed hierarchical graph information aggregation. We also propose a graph contrastive learning module to learn global theme-aware sentence representations. Extensive experiments on the PubMed and arXiv benchmark datasets prove the effectiveness of CHANGES and the importance of capturing hierarchical structure information in modeling scientific papers.

cs.CL↗
arXiv2023

The Magic of IF: Investigating Causal Reasoning Abilities in Large Language Models of Code

Causal reasoning, the ability to identify cause-and-effect relationship, is crucial in human thinking. Although large language models (LLMs) succeed in many NLP tasks, it is still challenging for them to conduct complex causal reasoning like abductive reasoning and counterfactual reasoning. Given the fact that programming code may express causal relations more often and explicitly with conditional statements like ``if``, we want to explore whether Code-LLMs acquire better causal reasoning abilities. Our experiments show that compared to text-only LLMs, Code-LLMs with code prompts are significantly better in causal reasoning. We further intervene on the prompts from different aspects, and discover that the programming structure is crucial in code prompt design, while Code-LLMs are robust towards format perturbations.

cs.CL↗
arXiv2023

Revisiting Parallel Context Windows: A Frustratingly Simple Alternative and Chain-of-Thought Deterioration

We identify two crucial limitations in the evaluation of recent parallel-integrated method Parallel Context Windows (PCW), which extends the maximum context lengths of language models, e.g., 2048 for LLaMA, by harnessing window-wise attention and positional embedding techniques. We first show that a simple yet strong baseline, weighted sum ensemble, is missing for the in-context few-shot classification. Moreover, on more challenging Chain-of-Thought (CoT) reasoning (e.g., HotpotQA), PCW would present unexpected deterioration regarding question miscomprehension and false inference. Based on our findings, we suggest that the existing PCW design may not guarantee sufficient improvement and practicality in handling lengthy documents in real-world applications. More community efforts on enabling language models' long context understanding ability should be paid.

cs.CL↗
arXiv2023

Hierarchical Adaptive Voxel-guided Sampling for Real-time Applications in Large-scale Point Clouds

While point-based neural architectures have demonstrated their efficacy, the time-consuming sampler currently prevents them from performing real-time reasoning on scene-level point clouds. Existing methods attempt to overcome this issue by using random sampling strategy instead of the commonly-adopted farthest point sampling~(FPS), but at the expense of lower performance. So the effectiveness/efficiency trade-off remains under-explored. In this paper, we reveal the key to high-quality sampling is ensuring an even spacing between points in the subset, which can be naturally obtained through a grid. Based on this insight, we propose a hierarchical adaptive voxel-guided point sampler with linear complexity and high parallelization for real-time applications. Extensive experiments on large-scale point cloud detection and segmentation tasks demonstrate that our method achieves competitive performance with the most powerful FPS, at an amazing speed that is more than 100 times faster. This breakthrough in efficiency addresses the bottleneck of the sampling step when handling scene-level point clouds. Furthermore, our sampler can be easily integrated into existing models and achieves a 20$\sim$80\% reduction in runtime with minimal effort. The code will be available at https://github.com/OuyangJunyuan/pointcloud-3d-detector-tensorrt

cs.CV↗
arXiv2023

Boosting Event Extraction with Denoised Structure-to-Text Augmentation

Event extraction aims to recognize pre-defined event triggers and arguments from texts, which suffer from the lack of high-quality annotations. In most NLP applications, involving a large scale of synthetic training data is a practical and effective approach to alleviate the problem of data scarcity. However, when applying to the task of event extraction, recent data augmentation methods often neglect the problem of grammatical incorrectness, structure misalignment, and semantic drifting, leading to unsatisfactory performances. In order to solve these problems, we propose a denoised structure-to-text augmentation framework for event extraction DAEE, which generates additional training data through the knowledge-based structure-to-text generation model and selects the effective subset from the generated data iteratively with a deep reinforcement learning agent. Experimental results on several datasets demonstrate that the proposed method generates more diverse text representations for event extraction and achieves comparable results with the state-of-the-art.

cs.CL↗
arXiv2023

DiffuSum: Generation Enhanced Extractive Summarization with Diffusion

Extractive summarization aims to form a summary by directly extracting sentences from the source document. Existing works mostly formulate it as a sequence labeling problem by making individual sentence label predictions. This paper proposes DiffuSum, a novel paradigm for extractive summarization, by directly generating the desired summary sentence representations with diffusion models and extracting sentences based on sentence representation matching. In addition, DiffuSum jointly optimizes a contrastive sentence encoder with a matching loss for sentence representation alignment and a multi-class contrastive loss for representation diversity. Experimental results show that DiffuSum achieves the new state-of-the-art extractive results on CNN/DailyMail with ROUGE scores of $44.83/22.56/40.56$. Experiments on the other two datasets with different summary lengths also demonstrate the effectiveness of DiffuSum. The strong performance of our framework shows the great potential of adapting generative models for extractive summarization. To encourage more following work in the future, we have released our codes at \url{https://github.com/hpzhang94/DiffuSum}

cs.CL↗
arXiv2023

Unsupervised Dense Retrieval Training with Web Anchors

In this work, we present an unsupervised retrieval method with contrastive learning on web anchors. The anchor text describes the content that is referenced from the linked page. This shows similarities to search queries that aim to retrieve pertinent information from relevant documents. Based on their commonalities, we train an unsupervised dense retriever, Anchor-DR, with a contrastive learning task that matches the anchor text and the linked document. To filter out uninformative anchors (such as ``homepage'' or other functional anchors), we present a novel filtering technique to only select anchors that contain similar types of information as search queries. Experiments show that Anchor-DR outperforms state-of-the-art methods on unsupervised dense retrieval by a large margin (e.g., by 5.3% NDCG@10 on MSMARCO). The gain of our method is especially significant for search and question answering tasks. Our analysis further reveals that the pattern of anchor-document pairs is similar to that of search query-document pairs. Code available at https://github.com/Veronicium/AnchorDR.

cs.IR↗
arXiv2023

Weighted Point Cloud Normal Estimation

Existing normal estimation methods for point clouds are often less robust to severe noise and complex geometric structures. Also, they usually ignore the contributions of different neighbouring points during normal estimation, which leads to less accurate results. In this paper, we introduce a weighted normal estimation method for 3D point cloud data. We innovate in two key points: 1) we develop a novel weighted normal regression technique that predicts point-wise weights from local point patches and use them for robust, feature-preserving normal regression; 2) we propose to conduct contrastive learning between point patches and the corresponding ground-truth normals of the patches' central points as a pre-training process to facilitate normal regression. Comprehensive experiments demonstrate that our method can robustly handle noisy and complex point clouds, achieving state-of-the-art performance on both synthetic and real-world datasets.

cs.CV↗
arXiv2023

Hierarchical Dialogue Understanding with Special Tokens and Turn-level Attention

Compared with standard text, understanding dialogue is more challenging for machines as the dynamic and unexpected semantic changes in each turn. To model such inconsistent semantics, we propose a simple but effective Hierarchical Dialogue Understanding model, HiDialog. Specifically, we first insert multiple special tokens into a dialogue and propose the turn-level attention to learn turn embeddings hierarchically. Then, a heterogeneous graph module is leveraged to polish the learned embeddings. We evaluate our model on various dialogue understanding tasks including dialogue relation extraction, dialogue emotion recognition, and dialogue act classification. Results show that our simple approach achieves state-of-the-art performance on all three tasks above. All our source code is publicly available at https://github.com/ShawX825/HiDialog.

cs.CL↗
arXiv2023

GraphMAE2: A Decoding-Enhanced Masked Self-Supervised Graph Learner

Graph self-supervised learning (SSL), including contrastive and generative approaches, offers great potential to address the fundamental challenge of label scarcity in real-world graph data. Among both sets of graph SSL techniques, the masked graph autoencoders (e.g., GraphMAE)--one type of generative method--have recently produced promising results. The idea behind this is to reconstruct the node features (or structures)--that are randomly masked from the input--with the autoencoder architecture. However, the performance of masked feature reconstruction naturally relies on the discriminability of the input features and is usually vulnerable to disturbance in the features. In this paper, we present a masked self-supervised learning framework GraphMAE2 with the goal of overcoming this issue. The idea is to impose regularization on feature reconstruction for graph SSL. Specifically, we design the strategies of multi-view random re-mask decoding and latent representation prediction to regularize the feature reconstruction. The multi-view random re-mask decoding is to introduce randomness into reconstruction in the feature space, while the latent representation prediction is to enforce the reconstruction in the embedding space. Extensive experiments show that GraphMAE2 can consistently generate top results on various public datasets, including at least 2.45% improvements over state-of-the-art baselines on ogbn-Papers100M with 111M nodes and 1.6B edges.

cs.LG↗
arXiv2023

Anomalous quantized plateaus in two-dimensional electron gas with gate confinement

Quantum information can be coded by the topologically protected edges of fractional quantum Hall (FQH) states. Investigation on FQH edges in the hope of searching and utilizing non-Abelian statistics has been a focused challenge for years. Manipulating the edges, e.g. to bring edges close to each other or to separate edges spatially, is a common and essential step for such studies. The FQH edge structures in a confined region are typically presupposed to be the same as that in the open region in analysis of experimental results, but whether they remain unchanged with extra confinement is obscure. In this work, we present a series of unexpected plateaus in a confined single-layer two-dimensional electron gas (2DEG), which are quantized at anomalous fractions such as 9/4, 17/11, 16/13 and the reported 3/2. We explain all the plateaus by assuming surprisingly larger filling factors in the confined region. Our findings enrich the understanding of edge states in the confined region and in the applications of gate manipulation, which is crucial for the experiments with quantum point contact and interferometer.

cond-mat.mes-hall↗
arXiv2023

DeepStruct: Pretraining of Language Models for Structure Prediction

We introduce a method for improving the structural understanding abilities of language models. Unlike previous approaches that finetune the models with task-specific augmentation, we pretrain language models on a collection of task-agnostic corpora to generate structures from text. Our structure pretraining enables zero-shot transfer of the learned knowledge that models have about the structure tasks. We study the performance of this approach on 28 datasets, spanning 10 structure prediction tasks including open information extraction, joint entity and relation extraction, named entity recognition, relation classification, semantic role labeling, event extraction, coreference resolution, factual probe, intent detection, and dialogue state tracking. We further enhance the pretraining with the task-specific training sets. We show that a 10B parameter language model transfers non-trivially to most tasks and obtains state-of-the-art performance on 21 of 28 datasets that we evaluate.

cs.CL↗