Papers by Abdalghani Abujabal

4 papers
Novel Feature Discovery for Task-Oriented Dialog Systems (2023.findings-eacl)

Copied to clipboard

Challenge: Prior work on novelty detection limits the scope of features represented by novel single intents to those represented by multiple user-perceived fine-grained features belonging to the same intent.
Approach: They propose to use a feature discovery technique to discover novel features from user utterances rather than single intent discovery to classify them into slots.
Outcome: The proposed approach consistently detects novel features from user utterances on two datasets.
Continuous Model Improvement for Language Understanding with Machine Translation (2021.naacl-industry)

Copied to clipboard

Challenge: a simple translation-test approach would fail the latency requirements of a live environment.
Approach: They show that annotating unlabeled utterances offline can improve performance . they demonstrate that an extrinsic evaluation can improve the performance if manual data is available .
Outcome: The proposed method improves performance in an extrinsic evaluation setting with real-world commercial dialog system in german.
Identifying and Resolving Annotation Changes for Natural Language Understanding (2021.naacl-industry)

Copied to clipboard

Challenge: Annotation conflict resolution is crucial for machine learning, says a new study . past work on annotation conflict resolution assumed data is collected at once . a a supervised neural model can resolve conflicts in data annotation but requires access to high-quality data .
Approach: They propose an approach to resolve annotation conflicts in a real-world context using a German dialog system.
Outcome: The proposed approach improves on a real-world dataset with 3.5M utterances in German.
ComQA: A Community-sourced Dataset for Complex Factoid Question Answering with Paraphrase Clusters (N19-1)

Copied to clipboard

Challenge: ComQA dataset captures question phenomena and the diverse ways in which they are formulated.
Approach: They propose a large dataset of real user questions that captures question phenomena and the diverse ways in which they are formulated.
Outcome: The proposed dataset can be a driver of future research on factoid question answering (QA).

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