Search results

Conversational agents are exploding in popularity. However, much work remains in the area of non goal-oriented conversations, despite significant growth in research interest over recent years. To advance the state of the art in conversational AI, Amazon launched the Alexa Prize, a 2.5-million dollar university competition where sixteen selected university teams built conversational agents to deliver the

Supplementing log filter-bank energies with i-vectors is a popular method for adaptive training of deep neural network acoustic models. While offline i-vectors (the target utterance or other relevant adaptation material is available for i-vector extraction prior to decoding) have been well studied, there is little analysis of online i-vectors and their robustness in multi-user scenarios where speaker changes

Dialog evaluation is a challenging problem, especially for non task-oriented dialogs where conversational success is not well-defined. We propose to evaluate dialog quality using topic-based metrics that describe the ability of a conversational bot to sustain coherent and engaging conversations on a topic, and the diversity of topics that a bot can handle. To detect conversation topics per utterance, we

In performance display advertising, bidders compete on behalf of advertisers for ad impressions, that is, the opportunity to display relevant ads on a publisher website. We consider bidding on behalf of online retailers who buy ad impressions hoping to realize value only from purchases attributed from clicks. The bidder has a two stage problem. In the first stage, the bidder has to select a small subset

Amazon is well-known for personalization and recommendations, which help customers discover items they might otherwise not have found. In this update to our original article, we discuss some of the changes as Amazon has grown.

We propose a novel latent factor model for recommending product size fits {Small, Fit, Large} to customers. Latent factors for customers and products in our model correspond to their physical true size, and are learnt from past product purchase and returns data. The outcome for a customer, product pair is predicted based on the difference between customer and product true sizes, and efficient algorithms

We introduce and describe the results of a novel shared task on bandit learning for machine translation. The task was organized jointly by Amazon and Heidelberg University for the first time at the Second Conference on Machine Translation (WMT 2017). The goal of the task is to encourage research on learning machine translation from weak user feedback instead of human references or post-edits. On each of

Learning a goal-oriented dialog policy is generally performed offline with supervised learning algorithms or online with reinforcement learning (RL). Additionally, as companies accumulate massive quantities of dialog transcripts between customers and trained human agents, encoder-decoder methods have gained popularity as agent utterances can be directly treated as supervision without the need for utterance-level

Learning attribute applicability of products in the Amazon catalog (e.g., predicting that a shoe should have a value for size, but not for battery-type) at scale is a challenge. The need for an interpretable model is contingent on (1) the lack of ground truth training data, (2) the need to utilise prior information about the underlying latent space and (3) the ability to understand the quality of predictions

We present a lightweight system to extract, store and manage metadata and provenance information of common artifacts in machine learning (ML) experiments: datasets, models, predictions, evaluations and training runs. Our system accelerates users in their ML workflow, and provides a basis for comparability and repeatability of ML experiments. We achieve this by tracking the lineage of produced artifacts

Optimization is commonly employed to determine the content of web pages, such as to maximize conversions on landing pages or click-through rates on search engine result pages. Often the layout of these pages can be decoupled into several separate decisions. For example, the composition of a landing page may involve deciding which image to show, which wording to use, what color background to display, etc

Randomized algorithms are central to modern machine learning. In the presence of massive datasets, researchers often turn to stochastic optimization to solve learning problems. Of particular interest is stochastic gradient descent (SGD), a first-order method that approximates the learning objective and gradient by a random point estimate. A classical question in learning theory is, if a randomized learner

Topics generated by topic models are typically presented as a list of topic terms. Automatic topic labelling is the task of generating a succinct label that summarises the theme or subject of a topic, with the intention of reducing the cognitive load of end-users when interpreting these topics. Traditionally, topic label systems focus on a single label modality, e.g. textual labels. In this work we propose

Bayesian optimization (BO) is a model-based approach for gradient-free black-box function optimization. Typically, BO is powered by a Gaussian process (GP), whose algorithmic complexity is cubic in the number of evaluations. Hence, GPbased BO cannot leverage large amounts of past or related function evaluations, for example, to warm start the BO procedure. We develop a multiple adaptive Bayesian linear

Generative adversarial networks (GANs) have recently seen a surge of interest in learning an inverse mapping—projecting the data back to the latent space. This learned mapping allows for image reconstruction—encoding and decoding—from a compact latent space. The choice of loss function(s) in this framework plays a critical role in determining the quality of the reconstruction. In this paper we explore possible

Interest in neural machine translation has grown rapidly as its effectiveness has been demonstrated across language and data scenarios. New research regularly introduces architectural and algorithmic improvements that lead to significant gains over “vanilla” NMT implementations. However, these new techniques are rarely evaluated in the context of previously published techniques, specifically those that

We present and apply two methods for addressing the problem of selecting relevant training data out of a general pool for use in tasks such as machine translation. Building on existing work on class-based language difference models [1], we first introduce a cluster-based method that uses Brown clusters to condense the vocabulary of the corpora. Secondly, we implement the cynical data selection method [2

The Moore-Lewis method of “intelligent selection of language model training data” is very effective, cheap, efficient... and also has structural problems. (1) The method defines relevance by playing language models trained on the in-domain and the out-of-domain (or data pool) corpora against each other. This powerful idea – which we set out to preserve – treats the two corpora as the opposing ends of a

We study the problem of a retailer that maximizes profit through joint replenishment, pricing and removal decisions. This problem is motivated by the observation that retailers usually retain rights to remove inventory from their network either by returning it to the suppliers or through liquidation in the face of random demand and capacity constraints. We develop a tractable dynamic program by leveraging

RubyStar is a dialog system designed to create “human-like” conversation by combining different response generation strategies. RubyStar conducts a non- task-oriented conversation on general topics by using an ensemble of rule-based, retrieval-based and generative methods. Topic detection, engagement monitoring, and context tracking are used for managing interaction. Predictable elements of conversation