Papers by Maarten Rijke

11 papers
Improving Multi-label Malevolence Detection in Dialogues through Multi-faceted Label Correlation Enhancement (2022.acl-long)

Copied to clipboard

Challenge: Current methods for detecting dialogue malevolence neglect label correlation.
Approach: They propose to crowdsource a multi-label dataset for detecting malevolent dialogue responses and a model with label correlation enhanced CRF to measure the correlation between malevolence and negative emotions.
Outcome: The proposed model outperforms the best performing baseline method on precision, recall, F1, and Jaccard score by 16.1%, 11.9%, 12.0%, and 6.1% on malevolence.
Real World Conversational Entity Linking Requires More Than Zero-Shots (2024.findings-acl)

Copied to clipboard

Challenge: Entity linking (EL) in conversations is a key component of many downstream tasks such as semantic search.
Approach: They propose to use Fandom and Wikipedia to evaluate EL models' ability to generalize to a new unfamiliar KB without prior training.
Outcome: The proposed evaluation framework and dataset are tailored to facilitate the study.
MELoRA: Mini-Ensemble Low-Rank Adapters for Parameter-Efficient Fine-Tuning (2024.acl-long)

Copied to clipboard

Challenge: Large language models (LLMs) are the default paradigm for natural language processing (NLP) as the models’ scale and the diversity of tasks increase, fine-tuning becomes infeasible.
Approach: They propose to freeze original pretrained weights and train a group of mini LoRAs with only a small number of parameters and reduce their rank by 8 times .
Outcome: The proposed model uses fewer trainable parameters while maintaining a higher rank, thereby offering improved performance potential.
CAUSE: Counterfactual Assessment of User Satisfaction Estimation in Task-Oriented Dialogue Systems (2024.findings-acl)

Copied to clipboard

Challenge: a number of studies have evaluated user satisfaction estimation in TOD systems . current benchmarks for user satisfaction estimates are highly skewed towards dialogues for which the user is satisfied.
Approach: They leverage large language models to generate satisfaction-aware counterfactual dialogues to augment original dialogues of a test collection.
Outcome: The proposed models show higher robustness to increase in dissatisfaction labels than fine-tuned models.
Bootstrapped Pre-training with Dynamic Identifier Prediction for Generative Retrieval (2024.findings-acl)

Copied to clipboard

Challenge: Existing methods for document retrieval rely on static document identifiers . experimental results show that generative retrieval is outperforms dense retrieval in document retrievals.
Approach: They propose a bootstrapped pre-training method that dynamically adjusts document identifiers during pre-train to accommodate the continuing memorization of the corpus.
Outcome: The proposed method significantly outperforms existing pre-training generative retrieval baselines and performs well even in zero-shot settings.
Table Question Answering for Low-resourced Indic Languages (2024.emnlp-main)

Copied to clipboard

Challenge: TableQA is the task of answering questions over tables of structured information, returning individual cells or tables as output.
Approach: They propose a fully automatic large-scale tableQA data generation process for low-resource languages with limited budget.
Outcome: The proposed method outperforms state-of-the-art LLMs on two Indic languages with no tableQA datasets and models on different aspects including mathematical reasoning capabilities and zero-shot cross-lingual transfer.
What Makes a Good and Useful Summary? Incorporating Users in Automatic Summarization Research (2022.naacl-main)

Copied to clipboard

Challenge: Existing research on automatic text summarization does not fully align with students’ needs.
Approach: They propose a survey methodology that can be used to investigate the needs of users of automatically generated summaries.
Outcome: The proposed method can be easily adjusted to investigate different user groups.
Pretraining Data Detection for Large Language Models: A Divergence-based Calibration Method (2024.emnlp-main)

Copied to clipboard

Challenge: Existing methods to detect text in training corpus are limited due to their low token probabilities.
Approach: They propose a method to calibrate token probabilities for pretraining data detection by using a divergence-based calibration method.
Outcome: The proposed method significantly outperforms existing methods on Chinese text on English-language benchmarks and patents.
KnowTuning: Knowledge-aware Fine-tuning for Large Language Models (2024.emnlp-main)

Copied to clipboard

Challenge: Large language models (LLMs) are a default solution for many natural language processing tasks.
Approach: They propose a knowledge-aware fine-tuning method to improve LLMs' knowledge awareness . they propose augmentation and comparison stages to improve accuracy and reliability .
Outcome: The proposed method generates more facts with less factual error rate under fine-grained facts evaluation.
Context Does Matter: Implications for Crowdsourced Evaluation Labels in Task-Oriented Dialogue Systems (2024.findings-naacl)

Copied to clipboard

Challenge: Existing studies suggest using only a portion of the dialogue context in the annotation process, but the impact of this limitation on label quality remains unexplored.
Approach: They propose to use large language models to summarize the dialogue context to provide a rich and short description of the dialogue and to examine the impact of doing so on the annotator’s performance.
Outcome: The proposed model reduces the context and produces higher quality ratings but introduces ambiguity in usefulness ratings.
The SIFo Benchmark: Investigating the Sequential Instruction Following Ability of Large Language Models (2024.findings-emnlp)

Copied to clipboard

Challenge: Current evaluation resources for instruction following focus on single task instructions, but the instruction sequences in these benchmarks often lack coherence.
Approach: They propose to evaluate models’ abilities to follow multiple instructions through sequential instruction following tasks using four tasks to assess different aspects of sequential instruction followed.
Outcome: The proposed benchmark outperforms open-source and closed-source models on four tasks assessing different aspects of sequential instruction following.

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