My research focuses on handling human languages (like English or Japanese) with computers, so called natural language processing. In particular, I am interested in machine learning approaches that are both linguistically motivated, and tailored to applications such as machine translation and speech recognition. You can find more details, research page or publications page.
I like developing, and have created open-source software which I have summarized here: software page. I also have had a chance to give tutorials and talks on subjects such as machine translation, neural networks, and language modeling, which are posted on my teaching page in the hope that people will find them useful.
- 2017-5-24: I gave a talk at the Machine Translation Marathon in the Americas on Softmax Alternatives for Neural MT.
- 2017-3-12: We'll be presenting three papers at EACL on cross-lingual word embeddings for language modeling, real-time translation with neural machine translation, and learning about syntax with recurrent neural network grammars. Adhi's paper on RNNGs won the Outstanding Paper Award, congratulations!
- 2017-3-5: I wrote a tutorial on neural machine translation and sequence to sequence models, which should be good for beginners who are just getting into the field.
- 2017-1-16: We wrote a technical report on DyNet, our neural network toolkit. We describe the inner workings, and do speed benchmarks against other popular toolkits such as TensorFlow, Theano, and Chainer. Take a look!
- 2017-1-5: We'll be holding the First Workshop on Neural Machine Translation at ACL2017. Lots of great speakers and hopefully an interesting panel discussion, so please join if you're interested!
- 2016-11-1: Our paper on learning lexicons from speech will receive the Best Short Paper Award at EMNLP!
- 2016-11-1: Four papers accepted to EMNLP on combining neural and n-gram language models, incorporating lexicons into neural machine translation, controlling output length in encoder-decoders, and learning lexicons from parallel speech.
- 2016-10-20: Our Japanese-English system for the Workshop on Asian Translation achieved the best results on the task!
- 2016-9-12: One paper accepted to Speech Communication on learning persuasive dialog systems using framing.
- 2016-8-10: One paper on continuous-space rule selection for syntax-based machine translation to be presented at ACL!
- 2016-6-28: I'll be giving an invited talk on automatic programming at the AIST artificial intelligence seminar.
- 2016-6-13: Two papers to be presented at NAACL-HLT on morphological inflection generation, and segment-based active learning for machine translation!
- 2016-5-30: I'll be giving an invited talk on language modeling at the TITECH/Ochadai NLP seminar.
- 2016-3-12: I'll be giving an invited talk at the IPSJ-ONE event sponsored by the Information Processing Society of Japan.
- 2016-3-6: A book (in Japanese) that I co-authored with Masato Hagiwara and Yoh Okuno on "Natural Language Processing: Fundamentals and Techniques" will be going on sale!
- 2016-3-3: Our Computational Linguistics survey on optimization for machine translation has been put online. Take a look if you want to know more about machine learning techniques for MT.
- 2016-2-16: I wrote a chapter on machine translation for the new Iwanami data science book series on natural language processing.
- 2016-2-11: I gave an invited talk at the Karlsruhe Institute of Technology on simultaneous speech translation.
- 2016-1-15: Paper accepted to the IEEE Transactions on Audio Speech and Language Processing on using the modulation spectrum for speech synthesis!
- 2016-1-13: Presented 3 papers at IWSDS on dialog-based deception detection, active learning for constructing dialog systems, and LSTM-based dialog state tracking.
- 2015-12-13: Presented 4 papers at ASRU on user-adaptive dialogue, sentence compression using LSTMs, affective communication, and our speech recognition system for the MGB challenge!
- 2015-12-2: Presented 4 papers at IWSLT on parser self-training for machine translation, learning lexicons from phoneme transcriptions, translation of prosodic emphasis, and the NAIST speech recognition system.
- 2015-11-9: Presented 2 papers at ASE on automatically generating pseudo-code from source code, and the corresponding tool, which can be found here!
- 2015-11-7: Giving a talk on Simultaneous Speech Translation at several universities: CMU, JHU, USC, and UMD.
- 2015-10-16: Our system at the 2nd workshop on Asian translation got 1st place in all 4 main language pairs! See our paper for details, and an analysis of why neural network reranking helps machine translation.
- 2015-10-8: A paper was accepted to TACL on semantic parsing of ambiguous input through paraphrasing and verification.
- 2015-10-1: Started a 4 months as a visiting scholar at CMU, working with Chris Dyer.