• Publications
  • Influence
Spectral learning of weighted automata
TLDR
A derivation of the spectral method for learning WFA that puts emphasis on providing intuitions on the inner workings of the method and does not assume a strong background in formal algebraic methods is presented.
Spectral Learning for Non-Deterministic Dependency Parsing
TLDR
This paper presents a learning algorithm that, like other spectral methods, is efficient and non-susceptible to local minima, and shows how this algorithm can be formulated as a technique for inducing hidden structure from distributions computed by forward-backward recursions.
Results of the Sequence PredIction ChallengE (SPiCe): a Competition on Learning the Next Symbol in a Sequence
TLDR
The Sequence PredIction ChallengE (SPiCe) is an on-line competition that took place between March and July 2016 and the aim was to submit a ranking of the 5 most probable symbols to be the next symbol of each prefix.
Atalaya at SemEval 2019 Task 5: Robust Embeddings for Tweet Classification
TLDR
This article describes its participation in HatEval, a shared task aimed at the detection of hate speech against immigrants and women, and trained robust task-oriented subword-aware embeddings and computed tweet representations using a weighted-averaging strategy.
Unsupervised Spectral Learning of WCFG as Low-rank Matrix Completion
TLDR
A spectral method is derived for unsupervised learning of Weighted Context Free Grammars by finding a Hankel matrix that has low rank and is linearly constrained to represent a function computed by inside-outside recursions.
PAC-Learning Unambiguous k, l-NTS <= Languages
TLDR
It is proved that k, l-UNTS≤ languages are also PAC-learnable under the same conditions as k,l-substitutable languages.
Atalaya at TASS 2018: Sentiment Analysis with Tweet Embeddings and Data Augmentation
TLDR
This work presents the participation as team Atalaya in the task of polarity classification of tweets, which followed standard techniques in preprocessing, representation and classification, and also explored some novel ideas.
Bounding the Maximal Parsing Performance of Non-Terminally Separated Grammars
TLDR
This paper develops methods to find upper bounds for the unlabeled F1 performance that any UWNTS grammar can achieve over a given treebank and defines a new metric that is NP-Hard but solvable with specialized software.
A Structured Listwise Approach to Learning to Rank for Image Tagging
TLDR
A bilinear compatibility function inspired from zero-shot learning that allows us to rank tags according to their relevance to the image content and proposes different “tags from captions” schemes meant to capture user attention and intra-user agreement in a simple and effective manner.
Upper Bounds for Unsupervised Parsing with Unambiguous Non-Terminally Separated Grammars
TLDR
This paper develops a method to find an upper bound for the unlabeled F1 performance that any UNTS grammar can achieve over a given tree-bank and shows that the F1 parsing score of any UNts grammar can not be beyond 82.2% when the gold treebank is the WSJ10 corpus.
...
1
2
...