Papers by Jonathan K. Kummerfeld

27 papers
Aligning AI Research with the Needs of Clinical Coding Workflows: Eight Recommendations Based on US Data Analysis and Critical Review (2025.acl-long)

Copied to clipboard

Challenge: Clinical coding is labour-intensive and error-prone, which has motivated research towards full automation of the process.
Approach: They propose to use AI to improve evaluation methods and propose new methods to assist clinical coders in their workflows.
Outcome: The proposed methods can be improved and improved on existing methods and the existing ones to assist coders in their workflows.
A Novel Workflow for Accurately and Efficiently Crowdsourcing Predicate Senses and Argument Labels (2020.findings-emnlp)

Copied to clipboard

Challenge: Prior attempts to develop crowdsourcing methods have either had low accuracy or required substantial expert annotation.
Approach: They propose a multi-stage crowd workflow that reduces expert involvement without sacrificing accuracy.
Outcome: The proposed method reduces expert effort by 4x, from 56% to 14% of cases.
Compositional Demographic Word Embeddings (2020.emnlp-main)

Copied to clipboard

Challenge: Word embeddings are usually derived from corpora containing text from many individuals . however, they cannot account for user-specific word preferences, such as using the same word in different ways across contexts.
Approach: They propose a new form of personalized word embeddings that use demographic-specific word representations derived compositionally from full or partial demographic information for a user.
Outcome: The proposed representations outperform generic representations on two English language tasks.
Quantifying and Avoiding Unfair Qualification Labour in Crowdsourcing (2021.acl-short)

Copied to clipboard

Challenge: Existing research suggests that crowd workers need to complete a substantial amount of poorly paid work to earn a fair wage.
Approach: They propose to use a qualification that requires workers to have completed a certain number of tasks to earn a fair wage.
Outcome: The proposed qualification reduces the burden on workers while still collecting high quality data.
Improving Text-to-SQL Evaluation Methodology (P18-1)

Copied to clipboard

Challenge: Current evaluations of text-to-SQL systems are limited by the way they divide data into training and test sets.
Approach: They propose to standardize and improve existing and new text-to-SQL datasets . they propose a template-based slot-filling baseline that cannot generalize to new queries .
Outcome: The proposed system is competitive with prior work on multiple datasets and can be used on training and test sets.
An Evaluation Dataset for Intent Classification and Out-of-Scope Prediction (D19-1)

Copied to clipboard

Challenge: Task-oriented dialog systems need to know when a query falls outside their range of supported intents.
Approach: They propose a dataset that includes queries that are out-of-scope and 150 intent classes over 10 domains.
Outcome: The proposed dataset includes queries that are out-of-scope, i.e., queries that do not fall into any of the system’s supported intents.
Inconsistencies in Crowdsourced Slot-Filling Annotations: A Typology and Identification Methods (2020.coling-main)

Copied to clipboard

Challenge: Standard slot-filling models train or finetune on large datasets of carefully-annotated data that is domain specific.
Approach: They propose automatic methods to identify inconsistencies in crowd-annotated data . a slot-filling model can extract the tokens "New York" as a TO LOCATION slot in a query .
Outcome: The proposed methods reveal inconsistencies in data, though there is scope for improvement.
Exploring the Value of Personalized Word Embeddings (2020.coling-main)

Copied to clipboard

Challenge: a subset of words belonging to specific psycholinguistic categories vary more in their representations across users . combining generic and personalized word embeddings yields the best performance .
Approach: They propose personalized word embeddings and compare their performance to generic ones . they show that personalized word representations can be leveraged for improved performance .
Outcome: The proposed model can be used for authorship attribution.
Factors Influencing the Surprising Instability of Word Embeddings (N18-1)

Copied to clipboard

Challenge: Word embeddings are low-dimensional, dense vector representations that capture semantic properties of words.
Approach: They examine the stability of word embeddings by examining their properties and analyzing their effects on downstream tasks.
Outcome: The results show that even high frequency words exhibit substantial instability, which can have implications for downstream tasks.
Exploring Self-Identified Counseling Expertise in Online Support Forums (2021.findings-acl)

Copied to clipboard

Challenge: Increasing number of people engage in online health forums, making it important to understand the quality of the advice they receive.
Approach: They examine the role of expertise in responses to help-seeking posts . they find that a classifier can distinguish between peer and self-identified mental health professionals' interactions .
Outcome: The findings show that experts' language use differs between groups, and that their comments engage the support-seeker further.
Improving Low Compute Language Modeling with In-Domain Embedding Initialisation (2020.emnlp-main)

Copied to clipboard

Challenge: Existing approaches to train language models on in-domain data are limited.
Approach: They propose to initialise and freeze in-domain embeddings to provide a useful representation of rare words in English . they find that the standard configuration is not optimal when rare words are present .
Outcome: The proposed approach improves language modeling by providing a useful representation of rare words in English.
Micromodels for Efficient, Explainable, and Reusable Systems: A Case Study on Mental Health (2021.findings-emnlp)

Copied to clipboard

Challenge: Existing statistical models are not explainable, struggle in low-resource scenarios and cannot be reused for multiple tasks.
Approach: They propose a micromodel architecture that embeds domain knowledge and provides explanations throughout the model’s decision process.
Outcome: The proposed model is validated on depression classification, PTSD classification, and suicidal risk assessment tasks.
SLATE: A Super-Lightweight Annotation Tool for Experts (P19-3)

Copied to clipboard

Challenge: a new annotation tool is designed to fill the niche of a lightweight interface for terminal users . current tools are built with direct manipulation via a Graphical User Interface (GUI) this approach is time-consuming and difficult to modify .
Approach: They propose a terminal-based annotation tool that supports multiple annotations . they use a text-based interface that uses almost the entire screen to display documents .
Outcome: The proposed tool is designed to fill the niche of a lightweight interface for users with a terminal-based workflow.
Iterative Feature Mining for Constraint-Based Data Collection to Increase Data Diversity and Model Robustness (2020.emnlp-main)

Copied to clipboard

Challenge: Recent work on dialog has found that crowdsourced data can have limited diversity as workers tend to write simple variations from prompts.
Approach: They propose a general approach for guiding workers to write more diverse text by iteratively constraining their writing.
Outcome: The proposed approach improves performance on dialog tasks and improves on existing datasets.
Outlier Detection for Improved Data Quality and Diversity in Dialog Systems (N19-1)

Copied to clipboard

Challenge: Existing methods to detect outliers in text have been neglected in NLP . outlier detection is a problem in dialog systems where text is often no more than a few sentences in length.
Approach: They propose a technique that uses sentence embeddings to detect outliers in short texts using neural sentence embeds and distance-based outlier detection.
Outcome: The proposed technique detects outliers in a corpus of short texts while generating highly diverse corpora that produce more robust intent classification and slot-filling models.
World Knowledge for Abstract Meaning Representation Parsing (L18-1)

Copied to clipboard

Challenge: Abstract Meaning Representation (AMR) parsers are based on annotated graphs, but there is still room for improvement .
Approach: They examine the role played by world knowledge in parsing errors in a state-of-the-art parser . they examine the effects of different types of world knowledge on parsers .
Outcome: The proposed model improves on multiple fine-grained metrics, including a 6% increase in named entity F-score, and provides insight into the potential of world knowledge for future work in Abstract Meaning Representation parsing.
Should I Trust You? Detecting Deception in Negotiations using Counterfactual RL (2025.findings-acl)

Copied to clipboard

Challenge: Future human-AI interaction tools can build on our methods for deception detection by triggering friction to give users a chance to interrogate suspicious proposals.
Approach: They propose to use CTRL-D to detect deception in a board game called Diplomacy . CTRL is a counterfactual RL that has a good recall and almost perfect precision . future tools could build on this to reevaluate trust in suspicious negotiations .
Outcome: The proposed method detects human deception with a high precision when compared to a Large Language Model approach that flags many true messages as deceptive.
Your Students Don’t Use LLMs Like You Wish They Did (2026.acl-long)

Copied to clipboard

Challenge: Educational NLP systems are evaluated using engagement metrics and satisfaction surveys . authors identify a fundamental misalignment between pedagogical design and student usage patterns .
Approach: They propose a computational framework for measuring behaviour in student-AI dialogue . they validate their framework by analysing 12,650 messages from four courses .
Outcome: The proposed metrics outperform surveys and satisfaction surveys on student-AI dialogues.
Analyzing the Surprising Variability in Word Embedding Stability Across Languages (2021.emnlp-main)

Copied to clipboard

Challenge: Word embeddings are powerful representations that form the foundation of many natural language processing architectures.
Approach: They explore word embedding stability in a wide range of languages to gain insight into their stability.
Outcome: The proposed results provide insights into word embedding stability in English and other languages.
A Large-Scale Corpus for Conversation Disentanglement (P19-1)

Copied to clipboard

Challenge: a dataset of 77,563 messages manually annotated with reply-structure graphs disentangles conversations and defines internal conversation structure.
Approach: They use a dataset of 77,563 messages manually annotated with reply-structure graphs to disentangle conversations and define internal conversation structure.
Outcome: The new dataset is 16 times larger than all previous datasets combined and includes adjudication of annotation disagreements and context.
Personalized Help for Optimizing Low-Skilled Users’ Strategy (2025.naacl-short)

Copied to clipboard

Challenge: a natural language agent generates moves and messages based on player intentions . a dozen games with novice and experienced players generate useful advice .
Approach: a team of researchers augment a natural language agent to generate move and message advice . they use a game to simulate the intentions of novice and experienced players .
Outcome: The enhanced agent generates move and message advice based on player intentions . the agent helps novices compete with experienced players and even surpass them .
Effective Crowdsourcing for a New Type of Summarization Task (N18-2)

Copied to clipboard

Challenge: Currently, summarization research focuses on summarizing the entire text, but in practice, readers are often interested in only one aspect of the document or conversation.
Approach: They propose a new task where the goal is to summarize a particular aspect of a document.
Outcome: The proposed task is based on a crowdsourced data collection workflow that allows users to collect high-quality summaries.
Using Paraphrases to Study Properties of Contextual Embeddings (2022.naacl-main)

Copied to clipboard

Challenge: Previously, paraphrases have been used to probe whether compositionality is accurately captured by BERT, but we believe they can be used to explore many other questions.
Approach: They propose to use paraphrases as a unique source of data to analyze contextualized embeddings, with a particular focus on BERT.
Outcome: The proposed analysis of paraphrases and paraphrase representations using the Paraphrase Database shows that BERT handles polysemous words, but different representations in many cases.
Less is More: Explainable and Efficient ICD Code Prediction with Clinical Entities (2025.acl-long)

Copied to clipboard

Challenge: Clinical coding is labor-intensive and prone to delays, leading to global backlogs.
Approach: They propose an approach that combines Named Entity Recognition (NER) and Assertion Classification (AC) to filter for clinically important content before supervised code prediction.
Outcome: The proposed approach reduces training time by over half on a standard evaluation dataset compared to current methods . it uses Named Entity Recognition (NER) and Assertion Classification (AC) to filter for clinically important content before supervised code prediction.
Leveraging Similar Users for Personalized Language Modeling with Limited Data (2022.acl-long)

Copied to clipboard

Challenge: Recent work suggests that personalized models are more accurate for individual users than one-size-fits-all solutions.
Approach: They propose a model trained on users that are similar to a new user to find similarity between new and existing users.
Outcome: The proposed model can predict what a user will write when they join a platform and not enough text is available.
Interactive Text-to-SQL Generation via Editable Step-by-Step Explanations (2023.emnlp-main)

Copied to clipboard

Challenge: Existing approaches to generate SQL from natural language are still making many mistakes . a new interaction mechanism allows users to edit a step-by-step explanation of a query to fix errors.
Approach: They propose a mechanism that allows users to edit a step-by-step explanation of a query to fix errors.
Outcome: The proposed approach can achieve better performance than multiple SOTA approaches on multiple datasets and 24 participants.
Data Collection for Dialogue System: A Startup Perspective (N18-3)

Copied to clipboard

Challenge: Developing dialogue systems such as Apple Siri and Google Now requires high quality training data but data collection with crowdsourcing is largely an open question.
Approach: They propose to use crowdsourcing to collect data for a user intent classification task in a dialogue system.
Outcome: The proposed method improves the quality of the collected data and the model performance on real user queries.

What is GenGO?

GenGO is an NLP powered publication search system. It currenctly indexes 30k+ papers from ACL Anthology, and implements multi-aspect summarization, semantic search, and more!

Information

About
Limitations