Paraphrase generation is an important problem in NLP, especially in question answering, information retrieval, information extraction, conversation systems, to name a few. In this paper, we address the problem of generating paraphrases automatically. Our proposed method is based on a combination of deep generative models (VAE) with sequence-to-sequence models (LSTM) to generate paraphrases, given an input sentence. Traditional VAEs when combined with recurrent neural networks can generate free text but they are not suitable for paraphrase generation for a given sentence. We address this problem by conditioning the both, encoder and decoder sides of VAE, on the original sentence, so that it can generate the given sentence's paraphrases. Unlike most existing models, our model is simple, modular and can generate multiple paraphrases, for a given sentence. Quantitative evaluation of the proposed method on a benchmark paraphrase dataset demonstrates its efficacy, and its performance improvement over the state-of-the-art methods by a significant margin, whereas qualitative human evaluation indicate that the generated paraphrases are well-formed, grammatically correct, and are relevant to the input sentence. Furthermore, we evaluate our method on a newly released question paraphrase dataset, and establish a new baseline for future research.

Paraphrase generation is an important problem in NLP, especially in question answering, information retrieval, information extraction, conversation systems, to name a few. In this paper, we address the problem of generating paraphrases automatically. Our proposed method is based on a combination of deep generative models (VAE) with sequence-to-sequence models (LSTM) to generate paraphrases, given an input sentence. Traditional VAEs when combined with recurrent neural networks can generate free text but they are not suitable for paraphrase generation for a given sentence. We address this problem by conditioning the both, encoder and decoder sides of VAE, on the original sentence, so that it can generate the given sentence's paraphrases. Unlike most existing models, our model is simple, modular and can generate multiple paraphrases, for a given sentence. Quantitative evaluation of the proposed method on a benchmark paraphrase dataset demonstrates its efficacy, and its performance improvement over the state-of-the-art methods by a significant margin, whereas qualitative human evaluation indicate that the generated paraphrases are well-formed, grammatically correct, and are relevant to the input sentence. Furthermore, we evaluate our method on a newly released question paraphrase dataset, and establish a new baseline for future research.

Connecting different text attributes associated with the same entity (conflation) is important in business data analytics since it could help merge two different tables in a database to provide a more comprehensive profile of an entity. However, the conflation task is challenging because two text strings that describe the same entity could be quite different from each other for reasons such as misspelling. It is therefore critical to develop a conflation model that is able to truly understand the semantic meaning of the strings and match them at the semantic level. To this end, we develop a character-level deep conflation model that encodes the input text strings from character level into finite dimension feature vectors, which are then used to compute the cosine similarity between the text strings. The model is trained in an end-to-end manner using back propagation and stochastic gradient descent to maximize the likelihood of the correct association. Specifically, we propose two variants of the deep conflation model, based on long-short-term memory (LSTM) recurrent neural network (RNN) and convolutional neural network (CNN), respectively. Both models perform well on a real-world business analytics dataset and significantly outperform the baseline bag-of-character (BoC) model.

Recently, there is an increasing need to share medical images for research purpose. In order to respect and preserve patient privacy, most of the medical images are de-identified with protected health information (PHI) before research sharing. Since manual de-identification is time-consuming and tedious, so an automatic de-identification system is necessary and helpful for the doctors to remove text from medical images. A lot of papers have been written about algorithms of text detection and extraction, however, little has been applied to de-identification of medical images. Since the de-identification system is designed for end-users, it should be effective, accurate and fast. This paper proposes an automatic system to detect and extract text from medical images for de-identification purposes, while keeping the anatomic structures intact. First, considering the text have a remarkable contrast with the background, a region variance based algorithm is used to detect the text regions. In post processing, geometric constraints are applied to the detected text regions to eliminate over-segmentation, e.g., lines and anatomic structures. After that, a region based level set method is used to extract text from the detected text regions. A GUI for the prototype application of the text detection and extraction system is implemented, which shows that our method can detect most of the text in the images. Experimental results validate that our method can detect and extract text in medical images with a 99% recall rate. Future research of this system includes algorithm improvement, performance evaluation, and computation optimization.

The rise of social media has made information exchange faster and easier among the people. However, in recent times, the use of offensive language has seen an upsurge in social media. The main challenge for a service provider is to correctly identify such offensive posts and take necessary action to monitor and control their spread. In this work, we try to address this problem by using sophisticated deep learning techniques like LSTM, Bidirectional LSTM and Bidirectional GRU. Our proposed approach solves 3 different Sub-tasks provided in the SemEval-2019 task 6 which incorporates identification of offensive tweets as well as their categorization. We obtain significantly better results in the leader-board for Sub-task B and decent results for Sub-task A and Subtask C validating the fact that the proposed models can be used for automating the offensive post-detection task in social media.