The workshop on Scholarly Document Processing (SDP) started in 2020 to accelerate research, inform policy, and educate the public on natural language processing for scientific text. The fifth iteration of the workshop, SDP 2025 was held at the 63rd Annual Meeting of the Association for Computational Linguistics (ACL 2025) in Vienna as a hybrid event. The workshop saw a great increase in interest, with 26 submissions, of which 11 were accepted for the research track. The program consisted of a research track, invited talks and four shared tasks: (1) SciHal25: Hallucination Detection for Scientific Content, (2) SciVQA: Scientific Visual Question Answering, (3) ClimateCheck: Scientific Factchecking of Social Media Posts on Climate Change, and (4) Software Mention Detection in Scholarly Publications (SOMD 25). In addition to the four shared task overview papers, 18 shared task reports were accepted. The program was geared towards NLP, information extraction, information retrieval, and data mining for scholarly documents, with an emphasis on identifying and providing solutions to open challenges.

In this paper, we explore the question of whether large language models can support cost-efficient information extraction from tables. We introduce schema-driven information extraction, a new task that transforms tabular data into structured records following a human-authored schema. To assess various LLM’s capabilities on this task, we present a benchmark comprised of tables from four diverse domains: machine learning papers, chemistry literature, material science journals, and webpages. We use this collection of annotated tables to evaluate the ability of open-source and API-based language models to extract information from tables covering diverse domains and data formats. Our experiments demonstrate that surprisingly competitive performance can be achieved without requiring task-specific pipelines or labels, achieving F1 scores ranging from 74.2 to 96.1, while maintaining cost efficiency. Moreover, through detailed ablation studies and analyses, we investigate the factors contributing to model success and validate the practicality of distilling compact models to reduce API reliance.

In this paper we present SynKB, an open-source, automatically extracted knowledge base of chemical synthesis protocols. Similar to proprietary chemistry databases such as Reaxsys, SynKB allows chemists to retrieve structured knowledge about synthetic procedures. By taking advantage of recent advances in natural language processing for procedural texts, SynKB supports more flexible queries about reaction conditions, and thus has the potential to help chemists search the literature for conditions used in relevant reactions as they design new synthetic routes. Using customized Transformer models to automatically extract information from 6 million synthesis procedures described in U.S. and EU patents, we show that for many queries, SynKB has higher recall than Reaxsys, while maintaining high precision. We plan to make SynKB available as an open-source tool; in contrast, proprietary chemistry databases require costly subscriptions.

We present VALET, a framework for rule-based information extraction written in Python. VALET departs from legacy approaches predicated on cascading finite-state transducers, instead offering direct support for mixing heterogeneous information–lexical, orthographic, syntactic, corpus-analytic–in a succinct syntax that supports context-free idioms. We show how a handful of rules suffices to implement sophisticated matching, and describe a user interface that facilitates exploration for development and maintenance of rule sets. Arguing that rule-based information extraction is an important methodology early in the development cycle, we describe an experiment in which a VALET model is used to annotate examples for a machine learning extraction model. While learning to emulate the extraction rules, the resulting model generalizes them, recognizing valid extraction targets the rules failed to detect.

We consider whether machine models can facilitate the human development of rule sets for information extraction. Arguing that rule-based methods possess a speed advantage in the early development of new extraction capabilities, we ask whether this advantage can be increased further through the machine facilitation of common recurring manual operations in the creation of an extraction rule set from scratch. Using a historical rule set, we reconstruct and describe the putative manual operations required to create it. In experiments targeting one key operation—the enumeration of words occurring in particular contexts—we simulate the process or corpus review and word list creation, showing that several simple interventions greatly improve recall as a function of simulated labor.

With the ever-increasing pace of research and high volume of scholarly communication, scholars face a daunting task. Not only must they keep up with the growing literature in their own and related fields, scholars increasingly also need to rebut pseudo-science and disinformation. These needs have motivated an increasing focus on computational methods for enhancing search, summarization, and analysis of scholarly documents. However, the various strands of research on scholarly document processing remain fragmented. To reach out to the broader NLP and AI/ML community, pool distributed efforts in this area, and enable shared access to published research, we held the 3rd Workshop on Scholarly Document Processing (SDP) at COLING as a hybrid event (https://sdproc.org/2022/). The SDP workshop consisted of a research track, three invited talks and five Shared Tasks: 1) MSLR22: Multi-Document Summarization for Literature Reviews, 2) DAGPap22: Detecting automatically generated scientific papers, 3) SV-Ident 2022: Survey Variable Identification in Social Science Publications, 4) SKGG: Scholarly Knowledge Graph Generation, 5) MuP 2022: Multi Perspective Scientific Document Summarization. The program was geared towards NLP, information retrieval, and data mining for scholarly documents, with an emphasis on identifying and providing solutions to open challenges.

Argument mining targets structures in natural language related to interpretation and persuasion which are central to scientific communication. Most scholarly discourse involves interpreting experimental evidence and attempting to persuade other scientists to adopt the same conclusions. While various argument mining studies have addressed student essays and news articles, those that target scientific discourse are still scarce. This paper surveys existing work in argument mining of scholarly discourse, and provides an overview of current models, data, tasks, and applications. We identify a number of key challenges confronting argument mining in the scientific domain, and suggest some possible solutions and future directions.

With the ever-increasing pace of research and high volume of scholarly communication, scholars face a daunting task. Not only must they keep up with the growing literature in their own and related fields, scholars increasingly also need to rebut pseudo-science and disinformation. These needs have motivated an increasing focus on computational methods for enhancing search, summarization, and analysis of scholarly documents. However, the various strands of research on scholarly document processing remain fragmented. To reach out to the broader NLP and AI/ML community, pool distributed efforts in this area, and enable shared access to published research, we held the 2nd Workshop on Scholarly Document Processing (SDP) at NAACL 2021 as a virtual event (https://sdproc.org/2021/). The SDP workshop consisted of a research track, three invited talks, and three Shared Tasks (LongSumm 2021, SCIVER, and 3C). The program was geared towards the application of NLP, information retrieval, and data mining for scholarly documents, with an emphasis on identifying and providing solutions to open challenges.

Next to keeping up with the growing literature in their own and related fields, scholars increasingly also need to rebut pseudo-science and disinformation. To address these challenges, computational work on enhancing search, summarization, and analysis of scholarly documents has flourished. However, the various strands of research on scholarly document processing remain fragmented. To reach to the broader NLP and AI/ML community, pool distributed efforts and enable shared access to published research, we held the 1st Workshop on Scholarly Document Processing at EMNLP 2020 as a virtual event. The SDP workshop consisted of a research track (including a poster session), two invited talks and three Shared Tasks (CL-SciSumm, Lay-Summ and LongSumm), geared towards easier access to scientific methods and results. Website: https://ornlcda.github.io/SDProc

We consider the problem of populating multi-part knowledge frames from textual information distributed over multiple sentences in a document. We present a corpus constructed by aligning papers from the cellular signaling literature to a collection of approximately 50,000 reference frames curated by hand as part of a decade-long project. We present and evaluate two approaches to the challenging problem of reconstructing these frames, which formalize biological assays described in the literature. One approach is based on classifying candidate records nominated by sentence-local entity co-occurrence. In the second approach, we introduce a novel virtual register machine traverses an article and generates frames, trained on our reference data. Our evaluations show that success in the task ultimately hinges on an integration of evidence spread across the discourse.

We describe a method for identifying and performing functional analysis of structured regions that are embedded in natural language documents, such as tables or key-value lists. Such regions often encode information according to ad hoc schemas and avail themselves of visual cues in place of natural language grammar, presenting problems for standard information extraction algorithms. Unlike previous work in table extraction, which assumes a relatively noiseless two-dimensional layout, our aim is to accommodate a wide variety of naturally occurring structure types. Our approach has three main parts. First, we collect and annotate a a diverse sample of “naturally” occurring structures from several sources. Second, we use probabilistic text segmentation techniques, featurized by skip bigrams over spatial and token category cues, to automatically identify contiguous regions of structured text that share a common schema. Finally, we identify the records and fields within each structured region using a combination of distributional similarity and sequence alignment methods, guided by minimal supervision in the form of a single annotated record. We evaluate the last two components individually, and conclude with a discussion of further work.

We discuss a named entity recognition system for Arabic, and show how we incorporated the information provided by MADA, a full morphological tagger which uses a morphological analyzer. Surprisingly, the relevant features used are the capitalization of the English gloss chosen by the tagger, and the fact that an analysis is returned (that a word is not OOV to the morphological analyzer). The use of the tagger also improves over a third system which just uses a morphological analyzer, yielding a 14\% reduction in error over the baseline. We conduct a thorough error analysis to identify sources of success and failure among the variations, and show that by combining the systems in simple ways we can significantly influence the precision-recall trade-off.

We survey the evaluation methodology adopted in Information Extraction (IE), as defined in the MUC conferences and in later independent efforts applying machine learning to IE. We point out a number of problematic issues that may hamper the comparison between results obtained by different researchers. Some of them are common to other NLP tasks: e.g., the difficulty of exactly identifying the effects on performance of the data (sample selection and sample size), of the domain theory (features selected), and of algorithm parameter settings. Issues specific to IE evaluation include: how leniently to assess inexact identification of filler boundaries, the possibility of multiple fillers for a slot, and how the counting is performed. We argue that, when specifying an information extraction task, a number of characteristics should be clearly defined. However, in the papers only a few of them are usually explicitly specified. Our aim is to elaborate a clear and detailed experimental methodology and propose it to the IE community. The goal is to reach a widespread agreement on such proposal so that future IE evaluations will adopt the proposed methodology, making comparisons between algorithms fair and reliable. In order to achieve this goal, we will develop and make available to the community a set of tools and resources that incorporate a standardized IE methodology.