Khanh Nguyen - Semantic Scholar

Skip to search formSkip to main contentSkip to account menuKhanh NguyenUniversity of California, Berkeleyhttp://khanhptnk.github.io (opens in a new tab)Publications15 h-index 9Citations744Highly Influential Citations41Follow Author...Author pages are created from data sourced from our academic… show morePublicationsCiting AuthorsReferenced AuthorsCo-AuthorsCo-AuthorHas PDFMore FiltersMore FiltersFiltersSort by Most Influential PapersSort by Citation CountSort by Recency

Help, Anna! Visual Navigation with Natural Multimodal Assistance via Retrospective Curiosity-Encouraging Imitation Learning

Khanh NguyenHal DauméComputer ScienceConference on Empirical Methods in Natural…
• 4 September 2019

TLDR“Help, Anna!” (HANNA), an interactive photo-realistic simulator in which an agent fulfills object-finding tasks by requesting and interpreting natural language-and-vision assistance, and an imitation learning algorithm that teaches the agent to avoid repeating past mistakes while simultaneously predicting its own chances of making future progress.Expand

• 164
• 12
• PDF (opens in a new tab)

ACL (opens in a new tab)SaveAlertCite

Reinforcement Learning for Bandit Neural Machine Translation with Simulated Human Feedback

Khanh NguyenHal DauméJ. Boyd-GraberComputer ScienceConference on Empirical Methods in Natural…
• 1 July 2017

TLDRA reinforcement learning algorithm that improves neural machine translation systems from simulated human feedback that combines the advantage actor-critic algorithm with the attention-based neural encoder-decoder architecture and effectively optimizes traditional corpus-level machine translation metrics.Expand

• 143
• 9
• PDF (opens in a new tab)

ACL (opens in a new tab)SaveAlertCite

Posterior calibration and exploratory analysis for natural language processing models

Khanh NguyenBrendan T. O'ConnorComputer Science, LinguisticsConference on Empirical Methods in Natural…
• 21 August 2015

TLDRIt is argued that the quality of a model' s posterior distribution can and should be directly evaluated, as to whether probabilities correspond to empirical frequencies, and NLP uncertainty can be projected not only to pipeline components, but also to exploratory data analysis, telling a user when to trust and not trust the NLP analysis.Expand

• 149
• 7
• PDF (opens in a new tab)

ACL (opens in a new tab)SaveAlertCite

Vision-Based Navigation With Language-Based Assistance via Imitation Learning With Indirect Intervention

Khanh NguyenDebadeepta DeyChris BrockettW. DolanComputer ScienceComputer Vision and Pattern Recognition
• 10 December 2018

TLDRTo model language-based assistance, a general framework termed Imitation Learning with Indirect Intervention (I3L) is developed, and a solution that is effective on the VNLA task is proposed that significantly improves the success rate of the learning agent over other baselines on both seen and unseen environments.Expand

• 144
• 5
• PDF (opens in a new tab)

IEEE (opens in a new tab)SaveAlertCite

Global Voices: Crossing Borders in Automatic News Summarization

Khanh NguyenHal DauméComputer Science, LinguisticsConference on Empirical Methods in Natural…
• 1 October 2019

TLDRThe effect of translation quality in cross-lingual summarization is studied, comparing a translate-then-summarize approach with several baselines, and the results highlight the limitations of the ROUGE metric that are overlooked in monolingual summarizations.Expand

• 38
• 2
• PDF (opens in a new tab)

ACL (opens in a new tab)SaveAlertCite

Interactive Learning from Activity Description

Khanh NguyenDipendra MisraR. SchapireMiroslav Dud'ikPatrick ShaftoComputer ScienceInternational Conference on Machine Learning
• 13 February 2021

TLDRA novel interactive learning protocol that enables training request-fulfilling agents by verbally describing their activities to achieve competitive success rates without requiring the teaching agent to be able to demonstrate the desired behavior using the learning agent's actions.Expand

• 35
• 2
• PDF (opens in a new tab)

arXiv (opens in a new tab)SaveAlertCite

The UMD machine translation systems at IWSLT 2015

Amittai AxelrodAhmed Elgohary+4 authors Marine CarpuatComputer Science, LinguisticsInternational Workshop on Spoken Language…
• 2015

TLDRThe University of Maryland machine translation systems submitted to the IWSLT 2015 French-English and Vietnamese-English tasks are described and novel data selection techniques to select relevant information from the large French- English training corpora are applied, and neural language models are tested.Expand

• 3
• 1
• PDF (opens in a new tab)

ACL (opens in a new tab)SaveAlertCite

Language-guided World Models: A Model-based Approach to AI Control

Alex ZhangKhanh NguyenJens TuylsAlbert LinKarthik NarasimhanComputer ScienceSPLUROBONLP
• 24 January 2024

TLDRLanguage-Guided World Models (LWMs) are developed, which can capture environment dynamics by reading language descriptions, allowing humans to simultaneously alter their behavior on multiple tasks with concise language feedback and enable agents to self-learn from texts originally written to instruct humans.Expand

• 10
• 1
• PDF (opens in a new tab)

ACL (opens in a new tab)SaveAlertCite

A Framework for Learning to Request Rich and Contextually Useful Information from Humans

Khanh NguyenYonatan BiskHal Daum'eComputer ScienceInternational Conference on Machine Learning
• 14 October 2021

TLDRA general interactive framework is presented that enables an agent to request and interpret rich, contextually useful information from an assistant that has knowledge about the task and the environment and demonstrates the practicality of the framework on a simulated human-assisted navigation problem.Expand

• 20
• 1
• PDF (opens in a new tab)

arXiv (opens in a new tab)SaveAlertCite

Probabilities of Chat LLMs Are Miscalibrated but Still Predict Correctness on Multiple-Choice Q&A

Benjamin PlautKhanh NguyenTu TrinhComputer ScienceTrans. Mach. Learn. Res.
• 20 February 2024

TLDRIt is shown that when models have the option to abstain, performance can be improved by selectively abstaining based on the MSP of the initial model response, using only a small amount of labeled data to choose the MSP threshold.Expand

• 10
• 1
• PDF (opens in a new tab)

arXiv (opens in a new tab)SaveAlertCite...12...

Co-Authors

• Hal Daumé
• 237 Publications • 22,095 Citations

• Hal Daum'e
• 19 Publications • 1,974 Citations

• Kianté Brantley
• 38 Publications • 969 Citations

• Jens Tuyls
• 13 Publications • 350 Citations

View All Co-AuthorsBy clicking accept or continuing to use the site, you agree to the terms outlined in our Privacy Policy (opens in a new tab), Terms of Service (opens in a new tab), and Dataset License (opens in a new tab)ACCEPT & CONTINUEor Only Accept Required