Writing

The Lifecycle View of Trustworthy AI

IBM Data Science in Practice (Medium) · October 22, 2021

Discusses the five pillars of trust—fairness, robustness, explainability, privacy, and transparency—and IBM Research's open-source AI toolkits.

NLU Enrichments Part 3: How to Utilize Metadata, Concepts, and Categories

Medium · January 5, 2018

Extracting meaning from unstructured data using metadata, category and concept enrichments.

NLU Enrichments Part 2: How to Utilize Entities, Relations, and Semantic Roles

Medium · January 5, 2018

Using Natural Language Understanding to identify specific people, places, organizations and their relationships.

NLU Enrichments Part 1: How to Utilize Sentiment, Keywords, and Emotion

Medium · January 5, 2018

How Watson Natural Language Understanding enrichments can uncover what people think and feel about a topic.

Learning to translate with products of novices

Transactions of the Association for Computational Linguistics · January 1, 2013

Research on machine translation using crowdsourced contributions from non-expert translators.

Nerit: Named entity recognition for informal text

Human Language Technology Center of Excellence, Johns Hopkins · January 1, 2013

Developed named entity recognition system optimized for informal text like social media.

Using targeted paraphrasing and monolingual crowdsourcing to improve translation

ACM Transactions on Intelligent Systems and Technology (TIST) · January 1, 2013

Research on leveraging crowdsourcing and paraphrasing techniques to improve machine translation quality.

The value of monolingual crowdsourcing in a real-world translation scenario

Proceedings of the Sixth Workshop on Statistical Machine Translation · January 1, 2011

Evaluated the effectiveness of monolingual crowdsourcing for improving translation in practical applications.

Using monolingual human computation to improve language translation via targeted paraphrase

HComp'10 Proceedings of the ACM SIGKDD Workshop on Human Computation · January 1, 2010

Workshop paper on using monolingual human computation for translation improvement through paraphrasing.

Crowdsourcing the evaluation of a domain-adapted named entity recognition system

Human Language Technologies (HLT) - Annual Conference of the North American Chapter of the ACL · January 1, 2010

Applied crowdsourcing methods to evaluate named entity recognition systems adapted to specific domains.

Error driven paraphrase annotation using mechanical turk

Proceedings of the NAACL HLT 2010 Workshop on Creating Speech and Language Data with Amazon's Mechanical Turk · January 1, 2010

Used crowdsourcing to collect paraphrase annotations driven by machine translation errors.

Improving translation via targeted paraphrasing

Proceedings of the 2010 Conference on Empirical Methods in Natural Language Processing (EMNLP) · January 1, 2010

Presented a method to improve machine translation by using targeted paraphrasing of source text.

Models of Synchronous Grammar Induction for SMT

Center for Language and Speech Processing, Johns Hopkins University · January 1, 2009

Research on synchronous grammar induction methods for statistical machine translation.