TACL 2019

CoQA: A Conversational Question Answering Challenge

TACL 2019 stanfordnlp/coqa-baselines

Humans gather information by engaging in conversations involving a series of interconnected questions and answers.

QUESTION ANSWERING READING COMPREHENSION

Synchronous Bidirectional Neural Machine Translation

TACL 2019 ZNLP/sb-nmt

In this paper, we introduce a synchronous bidirectional neural machine translation (SB-NMT) that predicts its outputs using left-to-right and right-to-left decoding simultaneously and interactively, in order to leverage both of the history and future information at the same time.

MACHINE TRANSLATION

What You Say and How You Say it: Joint Modeling of Topics and Discourse in Microblog Conversations

TACL 2019 zengjichuan/Topic_Disc

This paper presents an unsupervised framework for jointly modeling topic content and discourse behavior in microblog conversations.

Still a Pain in the Neck: Evaluating Text Representations on Lexical Composition

TACL 2019 vered1986/lexcomp

Building meaningful phrase representations is challenging because phrase meanings are not simply the sum of their constituent meanings.

WORD EMBEDDINGS

Learning Multilingual Word Embeddings in Latent Metric Space: A Geometric Approach

TACL 2019 anoopkunchukuttan/geomm

Our approach decouples learning the transformation from the source language to the target language into (a) learning rotations for language-specific embeddings to align them to a common space, and (b) learning a similarity metric in the common space to model similarities between the embeddings.

MULTILINGUAL WORD EMBEDDINGS

Semantic Neural Machine Translation using AMR

TACL 2019 freesunshine0316/semantic-nmt

It is intuitive that semantic representations can be useful for machine translation, mainly because they can help in enforcing meaning preservation and handling data sparsity (many sentences correspond to one meaning) of machine translation models.

MACHINE TRANSLATION

SECTOR: A Neural Model for Coherent Topic Segmentation and Classification

TACL 2019 sebastianarnold/SECTOR

From our extensive evaluation of 20 architectures, we report a highest score of 71. 6% F1 for the segmentation and classification of 30 topics from the English city domain, scored by our SECTOR LSTM model with bloom filter embeddings and bidirectional segmentation.

READING COMPREHENSION

Learning End-to-End Goal-Oriented Dialog with Maximal User Task Success and Minimal Human Agent Use

TACL 2019 IBM/modified-bAbI-dialog-tasks

In this work, we propose an end-to-end trainable method for neural goal-oriented dialog systems which handles new user behaviors at deployment by transferring the dialog to a human agent intelligently.

GOAL-ORIENTED DIALOG

GILE: A Generalized Input-Label Embedding for Text Classification

TACL 2019 idiap/gile

This forces their parametrization to be dependent on the label set size, and, hence, they are unable to scale to large label sets and generalize to unseen ones.

MULTI-TASK LEARNING TEXT CLASSIFICATION ZERO-SHOT LEARNING