In this paper we present two deep-learning systems that competed at SemEval-2017 Task 4 "Sentiment Analysis in Twitter". We participated in all subtasks for English tweets, involving message-level and topic-based sentiment polarity classification and quantification. We use Long Short-Term Memory (LSTM) networks augmented with two kinds of attention mechanisms, on top of word embeddings pre-trained on a big collection of Twitter messages. Also, we present a text processing tool suitable for social network messages, which performs tokenization, word normalization, segmentation and spell correction. Moreover, our approach uses no hand-crafted features or sentiment lexicons. We ranked 1 st (tie) in Subtask A, and achieved very competitive results in the rest of the Subtasks. Both the word embeddings and our text processing tool 1 are available to the research community.

A growing number of state-of-the-art transfer learning methods employ language models pretrained on large generic corpora. In this paper we present a conceptually simple and effective transfer learning approach that addresses the problem of catastrophic forgetting. Specifically, we combine the task-specific optimization function with an auxiliary language model objective, which is adjusted during the training process. This preserves language regularities captured by language models, while enabling sufficient adaptation for solving the target task. Our method does not require pretraining or finetuning separate components of the network and we train our models end-toend in a single step. We present results on a variety of challenging affective and text classification tasks, surpassing well established transfer learning methods with greater level of complexity.1

In this paper we present deep-learning models that submitted to the SemEval-2018 Task 1 competition: "Affect in Tweets". We participated in all subtasks for English tweets. We propose a Bi-LSTM architecture equipped with a multi-layer self attention mechanism. The attention mechanism improves the model performance and allows us to identify salient words in tweets, as well as gain insight into the models making them more interpretable. Our model utilizes a set of word2vec word embeddings trained on a large collection of 550 million Twitter messages, augmented by a set of word affective features. Due to the limited amount of task-specific training data, we opted for a transfer learning approach by pretraining the Bi-LSTMs on the dataset of Semeval 2017, Task 4A. The proposed approach ranked 1 st in Subtask E "Multi-Label Emotion Classification", 2 nd in Subtask A "Emotion Intensity Regression" and achieved competitive results in other subtasks.

The scarcity of large parallel corpora is an important obstacle for neural machine translation. A common solution is to exploit the knowledge of language models (LM) trained on abundant monolingual data. In this work, we propose a novel approach to incorporate a LM as prior in a neural translation model (TM). Specifically, we add a regularization term, which pushes the output distributions of the TM to be probable under the LM prior, while avoiding wrong predictions when the TM "disagrees" with the LM. This objective relates to knowledge distillation, where the LM can be viewed as teaching the TM about the target language. The proposed approach does not compromise decoding speed, because the LM is used only at training time, unlike previous work that requires it during inference. We present an analysis on the effects that different methods have on the distributions of the TM. Results on two low-resource machine translation datasets show clear improvements even with limited monolingual data.

Neural sequence-to-sequence models are currently the dominant approach in several natural language processing tasks, but require large parallel corpora. We present a sequenceto-sequence-to-sequence autoencoder (SEQ 3), consisting of two chained encoder-decoder pairs, with words used as a sequence of discrete latent variables. We apply the proposed model to unsupervised abstractive sentence compression, where the first and last sequences are the input and reconstructed sentences, respectively, while the middle sequence is the compressed sentence. Constraining the length of the latent word sequences forces the model to distill important information from the input. A pretrained language model, acting as a prior over the latent sequences, encourages the compressed sentences to be human-readable. Continuous relaxations enable us to sample from categorical distributions, allowing gradient-based optimization, unlike alternatives that rely on reinforcement learning. The proposed model does not require parallel text-summary pairs, achieving promising results in unsupervised sentence compression on benchmark datasets.