arXiv subjects

Chung-Wei Hang

Publications and source records attributed to Chung-Wei Hang.

14 recordsLinked to original sources

arXiv2026

What Really Matters for Table LLMs? A Meta-Evaluation of Model and Data Effects

Table modeling has progressed for decades. In this work, we revisit this trajectory and highlight emerging challenges in the LLM era, particularly the paradox of choice: the difficulty of attributing performance gains amid diverse base models and training sets in the context of table instruction tuning. We replicate four table LLMs by instruction-tuning three foundation models on four existing datasets, yielding 12 models. We then evaluate these models across 16 table benchmarks. Our study is the first to quantitatively disentangle the effects of training data and base model selection, revealing that base model choice plays a more dominant role than the training data itself. Generalization and reasoning remain challenging, inviting future effort on table modeling. Based on our findings, we share our thoughts on the future directions for table modeling.

cs.CL↗

arXiv2026

PRACTIQ: A Practical Conversational Text-to-SQL dataset with Ambiguous and Unanswerable Queries

Previous text-to-SQL datasets and systems have primarily focused on user questions with clear intentions that can be answered. However, real user questions can often be ambiguous with multiple interpretations or unanswerable due to a lack of relevant data. In this work, we construct a practical conversational text-to-SQL dataset called PRACTIQ, consisting of ambiguous and unanswerable questions inspired by real-world user questions. We first identified four categories of ambiguous questions and four categories of unanswerable questions by studying existing text-to-SQL datasets. Then, we generate conversations with four turns: the initial user question, an assistant response seeking clarification, the user's clarification, and the assistant's clarified SQL response with the natural language explanation of the execution results. For some ambiguous queries, we also directly generate helpful SQL responses, that consider multiple aspects of ambiguity, instead of requesting user clarification. To benchmark the performance on ambiguous, unanswerable, and answerable questions, we implemented large language model (LLM)-based baselines using various LLMs. Our approach involves two steps: question category classification and clarification SQL prediction. Our experiments reveal that state-of-the-art systems struggle to handle ambiguous and unanswerable questions effectively. We will release our code for data generation and experiments on GitHub.

cs.CL↗

arXiv2025

DSMentor: Enhancing Data Science Agents with Curriculum Learning and Online Knowledge Accumulation

Large language model (LLM) agents have shown promising performance in generating code for solving complex data science problems. Recent studies primarily focus on enhancing in-context learning through improved search, sampling, and planning techniques, while overlooking the importance of the order in which problems are tackled during inference. In this work, we develop a novel inference-time optimization framework, referred to as DSMentor, which leverages curriculum learning -- a strategy that introduces simpler task first and progressively moves to more complex ones as the learner improves -- to enhance LLM agent performance in challenging data science tasks. Our mentor-guided framework organizes data science tasks in order of increasing difficulty and incorporates a growing long-term memory to retain prior experiences, guiding the agent's learning progression and enabling more effective utilization of accumulated knowledge. We evaluate DSMentor through extensive experiments on DSEval and QRData benchmarks. Experiments show that DSMentor using Claude-3.5-Sonnet improves the pass rate by up to 5.2% on DSEval and QRData compared to baseline agents. Furthermore, DSMentor demonstrates stronger causal reasoning ability, improving the pass rate by 8.8% on the causality problems compared to GPT-4 using Program-of-Thoughts prompts. Our work underscores the importance of developing effective strategies for accumulating and utilizing knowledge during inference, mirroring the human learning process and opening new avenues for improving LLM performance through curriculum-based inference optimization.

cs.AI↗

arXiv2024

Towards a Holistic Evaluation of LLMs on Factual Knowledge Recall

Large language models (LLMs) have shown remarkable performance on a variety of NLP tasks, and are being rapidly adopted in a wide range of use cases. It is therefore of vital importance to holistically evaluate the factuality of their generated outputs, as hallucinations remain a challenging issue. In this work, we focus on assessing LLMs' ability to recall factual knowledge learned from pretraining, and the factors that affect this ability. To that end, we construct FACT-BENCH, a representative benchmark covering 20 domains, 134 property types, 3 answer types, and different knowledge popularity levels. We benchmark 31 models from 10 model families and provide a holistic assessment of their strengths and weaknesses. We observe that instruction-tuning hurts knowledge recall, as pretraining-only models consistently outperform their instruction-tuned counterparts, and positive effects of model scaling, as larger models outperform smaller ones for all model families. However, the best performance from GPT-4 still represents a large gap with the upper-bound. We additionally study the role of in-context exemplars using counterfactual demonstrations, which lead to significant degradation of factual knowledge recall for large models. By further decoupling model known and unknown knowledge, we find the degradation is attributed to exemplars that contradict a model's known knowledge, as well as the number of such exemplars. Lastly, we fine-tune LLaMA-7B in different settings of known and unknown knowledge. In particular, fine-tuning on a model's known knowledge is beneficial, and consistently outperforms fine-tuning on unknown and mixed knowledge. We will make our benchmark publicly available.

cs.CL↗

arXiv2023

UNITE: A Unified Benchmark for Text-to-SQL Evaluation

A practical text-to-SQL system should generalize well on a wide variety of natural language questions, unseen database schemas, and novel SQL query structures. To comprehensively evaluate text-to-SQL systems, we introduce a UNIfied benchmark for Text-to-SQL Evaluation (UNITE). It is composed of publicly available text-to-SQL datasets, containing natural language questions from more than 12 domains, SQL queries from more than 3.9K patterns, and 29K databases. Compared to the widely used Spider benchmark, we introduce $\sim$120K additional examples and a threefold increase in SQL patterns, such as comparative and boolean questions. We conduct a systematic study of six state-of-the-art (SOTA) text-to-SQL parsers on our new benchmark and show that: 1) Codex performs surprisingly well on out-of-domain datasets; 2) specially designed decoding methods (e.g. constrained beam search) can improve performance for both in-domain and out-of-domain settings; 3) explicitly modeling the relationship between questions and schemas further improves the Seq2Seq models. More importantly, our benchmark presents key challenges towards compositional generalization and robustness issues -- which these SOTA models cannot address well. Our code and data processing script are available at https://github.com/awslabs/unified-text2sql-benchmark

cs.CL↗

arXiv2023

Benchmarking Diverse-Modal Entity Linking with Generative Models

Entities can be expressed in diverse formats, such as texts, images, or column names and cell values in tables. While existing entity linking (EL) models work well on per modality configuration, such as text-only EL, visual grounding, or schema linking, it is more challenging to design a unified model for diverse modality configurations. To bring various modality configurations together, we constructed a benchmark for diverse-modal EL (DMEL) from existing EL datasets, covering all three modalities including text, image, and table. To approach the DMEL task, we proposed a generative diverse-modal model (GDMM) following a multimodal-encoder-decoder paradigm. Pre-training \Model with rich corpora builds a solid foundation for DMEL without storing the entire KB for inference. Fine-tuning GDMM builds a stronger DMEL baseline, outperforming state-of-the-art task-specific EL models by 8.51 F1 score on average. Additionally, extensive error analyses are conducted to highlight the challenges of DMEL, facilitating future research on this task.

cs.CL↗

arXiv2022

Importance of Synthesizing High-quality Data for Text-to-SQL Parsing

Recently, there has been increasing interest in synthesizing data to improve downstream text-to-SQL tasks. In this paper, we first examined the existing synthesized datasets and discovered that state-of-the-art text-to-SQL algorithms did not further improve on popular benchmarks when trained with augmented synthetic data. We observed two shortcomings: illogical synthetic SQL queries from independent column sampling and arbitrary table joins. To address these issues, we propose a novel synthesis framework that incorporates key relationships from schema, imposes strong typing, and conducts schema-distance-weighted column sampling. We also adopt an intermediate representation (IR) for the SQL-to-text task to further improve the quality of the generated natural language questions. When existing powerful semantic parsers are pre-finetuned on our high-quality synthesized data, our experiments show that these models have significant accuracy boosts on popular benchmarks, including new state-of-the-art performance on Spider.

cs.CL↗

arXiv2022

Improved Text Classification via Contrastive Adversarial Training

We propose a simple and general method to regularize the fine-tuning of Transformer-based encoders for text classification tasks. Specifically, during fine-tuning we generate adversarial examples by perturbing the word embeddings of the model and perform contrastive learning on clean and adversarial examples in order to teach the model to learn noise-invariant representations. By training on both clean and adversarial examples along with the additional contrastive objective, we observe consistent improvement over standard fine-tuning on clean examples. On several GLUE benchmark tasks, our fine-tuned BERT Large model outperforms BERT Large baseline by 1.7% on average, and our fine-tuned RoBERTa Large improves over RoBERTa Large baseline by 1.3%. We additionally validate our method in different domains using three intent classification datasets, where our fine-tuned RoBERTa Large outperforms RoBERTa Large baseline by 1-2% on average.

cs.CL↗

arXiv2021

Multilingual BERT Post-Pretraining Alignment

We propose a simple method to align multilingual contextual embeddings as a post-pretraining step for improved zero-shot cross-lingual transferability of the pretrained models. Using parallel data, our method aligns embeddings on the word level through the recently proposed Translation Language Modeling objective as well as on the sentence level via contrastive learning and random input shuffling. We also perform sentence-level code-switching with English when finetuning on downstream tasks. On XNLI, our best model (initialized from mBERT) improves over mBERT by 4.7% in the zero-shot setting and achieves comparable result to XLM for translate-train while using less than 18% of the same parallel data and 31% less model parameters. On MLQA, our model outperforms XLM-R_Base that has 57% more parameters than ours.

cs.CL↗

arXiv2020

Octa: Omissions and Conflicts in Target-Aspect Sentiment Analysis

Sentiments in opinionated text are often determined by both aspects and target words (or targets). We observe that targets and aspects interrelate in subtle ways, often yielding conflicting sentiments. Thus, a naive aggregation of sentiments from aspects and targets treated separately, as in existing sentiment analysis models, impairs performance. We propose Octa, an approach that jointly considers aspects and targets when inferring sentiments. To capture and quantify relationships between targets and context words, Octa uses a selective self-attention mechanism that handles implicit or missing targets. Specifically, Octa involves two layers of attention mechanisms for, respectively, selective attention between targets and context words and attention over words based on aspects. On benchmark datasets, Octa outperforms leading models by a large margin, yielding (absolute) gains in accuracy of 1.6% to 4.3%.

cs.CL↗

arXiv2018

Framing Matters: Predicting Framing Changes and Legislation from Topic News Patterns

News has traditionally been well researched, with studies ranging from sentiment analysis to event detection and topic tracking. We extend the focus to two surprisingly under-researched aspects of news: \emph{framing} and \emph{predictive utility}. We demonstrate that framing influences public opinion and behavior, and present a simple entropic algorithm to characterize and detect framing changes. We introduce a dataset of news topics with framing changes, harvested from manual surveys in previous research. Our approach achieves an F-measure of $F_1=0.96$ on our data, whereas dynamic topic modeling returns $F_1=0.1$. We also establish that news has \emph{predictive utility}, by showing that legislation in topics of current interest can be foreshadowed and predicted from news patterns.

cs.CY↗

arXiv2018

The Causal Link between News Framing and Legislation

We demonstrate that framing, a subjective aspect of news, is a causal precursor to both significant public perception changes, and to federal legislation. We posit, counter-intuitively, that topic news volume and mean article similarity increase and decrease together. We show that specific features of news, such as publishing volume , are predictive of both sustained public attention, measured by annual Google trend data, and federal legislation. We observe that public attention changes are driven primarily by periods of high news volume and mean similarity, which we call \emph{prenatal periods}. Finally, we demonstrate that framing during prenatal periods may be characterized by high-utility news \emph{keywords}.

cs.CY↗

arXiv2015

Empirical Study on Deep Learning Models for Question Answering

In this paper we explore deep learning models with memory component or attention mechanism for question answering task. We combine and compare three models, Neural Machine Translation, Neural Turing Machine, and Memory Networks for a simulated QA data set. This paper is the first one that uses Neural Machine Translation and Neural Turing Machines for solving QA tasks. Our results suggest that the combination of attention and memory have potential to solve certain QA problem.

cs.CL↗

arXiv2014

A Probabilistic Approach for Maintaining Trust Based on Evidence

Leading agent-based trust models address two important needs. First, they show how an agent may estimate the trustworthiness of another agent based on prior interactions. Second, they show how agents may share their knowledge in order to cooperatively assess the trustworthiness of others. However, in real-life settings, information relevant to trust is usually obtained piecemeal, not all at once. Unfortunately, the problem of maintaining trust has drawn little attention. Existing approaches handle trust updates in a heuristic, not a principled, manner. This paper builds on a formal model that considers probability and certainty as two dimensions of trust. It proposes a mechanism using which an agent can update the amount of trust it places in other agents on an ongoing basis. This paper shows via simulation that the proposed approach (a) provides accurate estimates of the trustworthiness of agents that change behavior frequently; and (b) captures the dynamic behavior of the agents. This paper includes an evaluation based on a real dataset drawn from Amazon Marketplace, a leading e-commerce site.

cs.MA↗