"What is the value of {templates}?" Rethinking Document Information Extraction Datasets for LLMs

• URL: http://arxiv.org/abs/2410.15484v1
• Date: Sun, 20 Oct 2024 19:42:30 GMT
• Title: "What is the value of {templates}?" Rethinking Document Information Extraction Datasets for LLMs
• Authors: Ran Zmigrod, Pranav Shetty, Mathieu Sibue, Zhiqiang Ma, Armineh Nourbakhsh, Xiaomo Liu, Manuela Veloso,
• Abstract summary: We present K2Q, a collection of five datasets converted from KIE to a prompt-response format using a plethora of bespoke templates. We empirically compare the performance of seven baseline generative models on K2Q with zero-shot prompting. We find that creating diverse and intricate KIE questions enhances the performance and robustness of VRDU models.
• Score: 19.07429412219697
• License:
• Abstract: The rise of large language models (LLMs) for visually rich document understanding (VRDU) has kindled a need for prompt-response, document-based datasets. As annotating new datasets from scratch is labor-intensive, the existing literature has generated prompt-response datasets from available resources using simple templates. For the case of key information extraction (KIE), one of the most common VRDU tasks, past work has typically employed the template "What is the value for the {key}?". However, given the variety of questions encountered in the wild, simple and uniform templates are insufficient for creating robust models in research and industrial contexts. In this work, we present K2Q, a diverse collection of five datasets converted from KIE to a prompt-response format using a plethora of bespoke templates. The questions in K2Q can span multiple entities and be extractive or boolean. We empirically compare the performance of seven baseline generative models on K2Q with zero-shot prompting. We further compare three of these models when training on K2Q versus training on simpler templates to motivate the need of our work. We find that creating diverse and intricate KIE questions enhances the performance and robustness of VRDU models. We hope this work encourages future studies on data quality for generative model training.

Related papers

• Rethinking Data Synthesis: A Teacher Model Training Recipe with Interpretation [8.013158752919722]
  Recent advances in large language model (LLM) training have highlighted the need for diverse, high-quality instruction data. We propose a paradigm shift named textbfNOMAD by investigating how to specifically train models for data generation.
  arXiv  Detail & Related papers  (2024-10-27T07:38:39Z)
• Less is More: Making Smaller Language Models Competent Subgraph Retrievers for Multi-hop KGQA [51.3033125256716]
  We model the subgraph retrieval task as a conditional generation task handled by small language models. Our base generative subgraph retrieval model, consisting of only 220M parameters, competitive retrieval performance compared to state-of-the-art models. Our largest 3B model, when plugged with an LLM reader, sets new SOTA end-to-end performance on both the WebQSP and CWQ benchmarks.
  arXiv  Detail & Related papers  (2024-10-08T15:22:36Z)
• CRAFT Your Dataset: Task-Specific Synthetic Dataset Generation Through Corpus Retrieval and Augmentation [51.2289822267563]
  We propose Corpus Retrieval and Augmentation for Fine-Tuning (CRAFT), a method for generating synthetic datasets. We use large-scale public web-crawled corpora and similarity-based document retrieval to find other relevant human-written documents. We demonstrate that CRAFT can efficiently generate large-scale task-specific training datasets for four diverse tasks.
  arXiv  Detail & Related papers  (2024-09-03T17:54:40Z)
• Jaeger: A Concatenation-Based Multi-Transformer VQA Model [0.13654846342364307]
  Document-based Visual Question Answering poses a challenging task between linguistic sense disambiguation and fine-grained multimodal retrieval. We propose Jaegar, a concatenation-based multi-transformer VQA model. Our approach has the potential to amplify the performance of these models through concatenation.
  arXiv  Detail & Related papers  (2023-10-11T00:14:40Z)
• Unlocking Model Insights: A Dataset for Automated Model Card Generation [4.167070553534516]
  We introduce a dataset of 500 question-answer pairs for 25 ML models. We employ annotators to extract the answers from the original paper. Our experiments with ChatGPT-3.5, LLaMa, and Galactica showcase a significant gap in the understanding of research papers by these LMs.
  arXiv  Detail & Related papers  (2023-09-22T04:46:11Z)
• AnnoLLM: Making Large Language Models to Be Better Crowdsourced Annotators [98.11286353828525]
  GPT-3.5 series models have demonstrated remarkable few-shot and zero-shot ability across various NLP tasks. We propose AnnoLLM, which adopts a two-step approach, explain-then-annotate. We build the first conversation-based information retrieval dataset employing AnnoLLM.
  arXiv  Detail & Related papers  (2023-03-29T17:03:21Z)
• When Can Models Learn From Explanations? A Formal Framework for Understanding the Roles of Explanation Data [84.87772675171412]
  We study the circumstances under which explanations of individual data points can improve modeling performance. We make use of three existing datasets with explanations: e-SNLI, TACRED, SemEval.
  arXiv  Detail & Related papers  (2021-02-03T18:57:08Z)
• KGPT: Knowledge-Grounded Pre-Training for Data-to-Text Generation [100.79870384880333]
  We propose a knowledge-grounded pre-training (KGPT) to generate knowledge-enriched text. We adopt three settings, namely fully-supervised, zero-shot, few-shot to evaluate its effectiveness. Under zero-shot setting, our model achieves over 30 ROUGE-L on WebNLG while all other baselines fail.
  arXiv  Detail & Related papers  (2020-10-05T19:59:05Z)
• Dense-Caption Matching and Frame-Selection Gating for Temporal Localization in VideoQA [96.10612095576333]
  We propose a video question answering model which effectively integrates multi-modal input sources and finds the temporally relevant information to answer questions. Our model is also comprised of dual-level attention (word/object and frame level), multi-head self-cross-integration for different sources (video and dense captions), and which pass more relevant information to gates. We evaluate our model on the challenging TVQA dataset, where each of our model components provides significant gains, and our overall model outperforms the state-of-the-art by a large margin.
  arXiv  Detail & Related papers  (2020-05-13T16:35:27Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.