Interpretation and diagnosis of machine learning models have gained renewed interest in recent years with breakthroughs in new approaches. We present Manifold, a framework that utilizes visual analysis techniques to support interpretation, debugging, and comparison of machine learning models in a more transparent and interactive manner. Conventional techniques usually focus on visualizing the internal logic of a specific model type (i.e., deep neural networks), lacking the ability to extend to a more complex scenario where different model types are integrated. To this end, Manifold is designed as a generic framework that does not rely on or access the internal logic of the model and solely observes the input (i.e., instances or features) and the output (i.e., the predicted result and probability distribution). We describe the workflow of Manifold as an iterative process consisting of three major phases that are commonly involved in the model development and diagnosis process: inspection (hypothesis), explanation (reasoning), and refinement (verification). The visual components supporting these tasks include a scatterplot-based visual summary that overviews the models' outcome and a customizable tabular view that reveals feature discrimination. We demonstrate current applications of the framework on the classification and regression tasks and discuss other potential machine learning use scenarios where Manifold can be applied.

Today recommenders are commonly used with various purposes, especially dealing with e-commerce and information filtering tools. Content-based recommenders rely on the concept of similarity between the bought/searched/visited item and all the items stored in a repository. It is a common belief that the user is interested in what issimilar to what she has already bought/searched/visited. We believe that there are some contexts in which this assumption is wrong: it is the case of acquiring unsearched but still useful items or pieces of information. This is called serendipity. Our purpose is to stimulate users and facilitate these serendipitous encounters to happen. This paper presents the design and implementation of a hybrid recommender system that joins a contentbased approach and serendipitous heuristics in order to mitigate the over-specialization problem with surprising suggestions. Background and MotivationInformation overload is a common issue among the modern information society. Information Filtering (IF) is a kind of intelligent computing techniques that mitigates this problem by providing the user with the most relevant information with respect to her information needs.Recommender systems (RSs) adopt IF techniques in order to provide customized information access for targeted domains.They can be viewed as intelligent systems that take input directly or indirectly from users and, based on their needs, preferences and usage patterns, provide personalized advices about products or services and can help people to filter useful information.Several definitions of RS have been given. According to [3]: "Recommender systems have the effect of guiding the user in a personalized way to interesting or useful objects in a large space of possible options". This definition makes it clear that user oriented guidance is critical in a RS.Among different recommendation techniques proposed in the literature, the content-based and the collaborative filtering approaches are the most widely adopted to date. Systems implementing the contentbased recommendation approach analyze a set of documents, usually textual descriptions of the items previously rated by an individual user, and build a model or profile of user interests based on the features of the objects rated by that user [14]. The profile is exploited to recommend new items of interest. Collaborative recommenders differ from content-based ones in that user opinions are used instead of content. They gather ratings about objects by users and store them in a centralized or distributed database. To provide user X with recommendations, the system computes the neighborhood of that user, i.e. the subset of users that have a taste similar to X. Similarity in taste is computed based on the similarity of ratings for objects that were rated by both users. The system then recommends objects that users in X's neighborhood indicated to like, provided that they have not yet been rated by X. Each type of filtering methods has its own weaknesses and strengths.In particular, the content-ba...

This work introduces Focused-Variation Network (FVN), a novel model to control language generation. The main problems in previous controlled language generation models range from the difficulty of generating text according to the given attributes, to the lack of diversity of the generated texts. FVN addresses these issues by learning disjoint discrete latent spaces for each attribute inside codebooks, which allows for both controllability and diversity, while at the same time generating fluent text. We evaluate FVN on two text generation datasets with annotated content and style, and show state-of-the-art performance as assessed by automatic and human evaluations.

No abstract

The quality of automatic speech recognition (ASR) is critical to Dialogue Systems as ASR errors propagate to and directly impact downstream tasks such as language understanding (LU). In this paper, we propose multi-task neural approaches to perform contextual language correction on ASR outputs jointly with LU to improve the performance of both tasks simultaneously. To measure the effectiveness of this approach we used a public benchmark, the 2nd Dialogue State Tracking (DSTC2) corpus. As a baseline approach, we trained task specific Statistical Language Models (SLM) and fine-tuned state-of-the-art Generalized Pre-training (GPT) Language Model to re-rank the n-best ASR hypotheses, followed by a model to identify the dialog act and slots. i) We further trained ranker models using GPT and Hierarchical CNN-RNN models with discriminatory losses to detect the best output given n-best hypotheses. We extended these ranker models to first select the best ASR output and then identify the dialogue act and slots in an end to end fashion. ii) We also proposed a novel joint ASR error correction and LU model, a word confusion pointer network (WCN-Ptr) with multihead self attention on top, which consumes the word confusions populated from the n-best. We show that the error rates of off the shelf ASR and following LU systems can be reduced significantly by 14% relative with joint models trained using small amounts of in-domain data.