Phong Le (Lê Phong in Vietnamese)

Amazon Alexa, Cambridge, UK

l[firstname][at]amazon.com

I'm an applied scientist at Amazon Alexa. Before, I was research fellow at the University of Manchester. I did a postdoc with Ivan Titov at the University of Amsterdam and the University of Edinburgh. I got a PhD from the University of Amsterdam under the supervision of (Jelle) Willem Zuidema.

I'm interested in neural networks and deep learning. My current work is to employ them to solve natural language processing tasks such as entity linking, coreference resolution, and dependency parsing. I'm also interested in semantic parsing, which is the topic of my master thesis.

Checkout my CV, Google Scholar, and github.

Publications

Natural Language Processing

2021

  • T. Tran, P. Le, S. Ananiadou. One-shot to Weakly-Supervised Relation Classification using Language Models. AKBC [pdf, bib, code]

2020

  • P. Le, W. Zuidema. DoLFIn: Distributions over Latent Features for Interpretability. COLING [pdf, bib, code, slides]

  • T. Tran, P. Le, S. Ananiadou. Revisiting Unsupervised Relation Extraction. ACL [pdf, bib, code]

2019

  • P. Le, I. Titov. Boosting Entity Linking Performance by Leveraging Unlabeled Documents. ACL (Best Paper Award nomination) [pdf, bib, code, slides]

  • P. Le, I. Titov. Distant Learning for Entity Linking with Automatic Noise Detection. ACL [pdf, bib, code, poster]

  • W. Zuidema, P. Le. Vector-based and Neural Models of Semantics (book chapter). In: Peter Hagoort (ed.), Human Language: From Genes and Brains to Behavior, MIT Press.

2018

  • P. Le, I. Titov. Improving Entity Linking by Modeling Latent Relations between Mentions. ACL. [pdf, bib, code]

2017

  • P. Le, I. Titov. Optimizing Differentiable Relaxations of Coreference Evaluation Metrics. CoNLL (oral). [pdf, bib, code, slides]

2016

  • P. Le, M. Dymetman, J-M. Renders. LSTM-based Mixture-of-Experts for Knowledge-Aware Dialogues. ACL Workshop on Representation Learning for NLP, 2016. [pdf, bib].

  • P. Le and W. Zuidema. Quantifying the vanishing gradient and long distance dependency problem in recursive neural networks and recursive LSTMs. ACL Workshop on Representation Learning for NLP, 2016. [pdf, bib]

2015

  • P. Le and W. Zuidema. The Forest Convolutional Network: Compositional Distributional Semantics with a Neural Chart and without Binarization. EMNLP [pdf, bib]

  • P. Le. Enhancing the Inside-Outside Recursive Neural Network Reranker for Dependency Parsing. International Conference on Parsing Technology (IWPT) (short paper) [pdf, bib]

  • P. Le and W. Zuidema. Compositional Distributional Semantics with Long Short Term Memory. Joint Conference on Lexical and Computational Semantics (*SEM) [pdf, bib, code]

  • P. Le and W. Zuidema. Unsupervised Dependency Parsing: Let's Use Supervised Parsers. NAACL-HLT [pdf, bib]

2014

  • P. Le and W. Zuidema. Inside-Outside Semantics: A Framework for Neural Models of Semantic Composition. NIPS 2014 Workshop on Deep Learning and Representation Learning [pdf]

  • P. Le and W. Zuidema. The Inside-Outside Recursive Neural Network model for Dependency Parsing. EMNLP [pdf, bib, code]

2013

  • P. Le, W. Zuidema, and R.J.H. Scha. Learning from errors: Using vector-based compositional semantics for parse reranking. ACL Workshop on Continuous Vector Space Models and their Compositionality (oral) [pdf, bib]

2012

  • P. Le and W. Zuidema. Learning Compositional Semantics for Open Domain Semantic Parsing. COLING [pdf, bib]


Computer Vision

  • P. Le, A.D. Duong, H.Q. Vu, and N.T. Pham (2009). Adaptive hybrid mean shift and particle filter. In proceedings of International Conference on Computing and Communication Technologies, 2009. RIVF’09, pages 1–4. [pdf]

  • P. Le, N.T. Pham, A.D. Duong, and H.Q. Vu (2008). Tracking a variable number of humans and motorcycles in video via Probability Hypothesis Density filter. IEEE Student Paper Contest. Seoul, Korea.


Theses

  • P. Le. Learning Vector Representations for Sentences - The Recursive Deep Learning Approach. PhD's Thesis, University of Amsterdam. [pdf]

  • P. Le. Learning Semantic Parsing. Master's Thesis, University of Amsterdam. [pdf] (STIL Thesis Prize)


Patent

  • P. Le, M. Dymetman, J. Renders. Dialog device with dialog support generated using a mixture of language models combined using a recurrent neural network. US Patent 20170316775, 2017 [pdf]


Academic services

  • Frequent reviewer for NLP conferences (ACL, EMNLP, NAACL, CoNLL, EACL, AACL, Coling), ML conferences (NeurIPS, ICML, ICLR), and AI conferences (AAAI) since 2016.

  • Administration chair for CoNLL 2018

  • Publication chair for *SEM 2016