In e-commerce portals, generating answers for product-related questions has become a crucial task. In this paper, we propose the task of product-aware answer generation, which tends to generate an accurate and complete answer from large-scale unlabeled e-commerce reviews and product attributes. Unlike existing question-answering problems, answer generation in e-commerce confronts three main challenges: (1) Reviews are informal and noisy; (2) joint modeling of reviews and key-value product attributes is challenging; (3) traditional methods easily generate meaningless answers. To tackle above challenges, we propose an adversarial learning based model, named PAAG, which is composed of three components: a question-aware review representation module, a key-value memory network encoding attributes, and a recurrent neural network as a sequence generator. Specifically, we employ a convolutional discriminator to distinguish whether our generated answer matches the facts. To extract the salience part of reviews, an attention-based review reader is proposed to capture the most relevant words given the question. Conducted on a large-scale real-world e-commerce dataset, our extensive experiments verify the effectiveness of each module in our proposed model. Moreover, our experiments show that our model achieves the state-of-the-art performance in terms of both automatic metrics and human evaluations.

As data people, we very typically spend a great deal of time summarizing our findings to stakeholders in a clear, concise and impactful way. Often times, due to the lack of infrastructure, we end up using presentation files with chart images. This can become a real pain when we need to make modifications or when the analysis needs to "live on". Typically, this is where BI (business intelligence) or dashboard tools shine. Unfortunately, this can be a major stumbling block for smaller shops who rely on a lot of local analysis and may not have the budget for a BI tool.

Google voice search is an omnipresent choice for users on the move via their mobile devices. Voice search has made its way into our everyday life and the trends suggest that it's here to stay. Voice search is a function that makes it possible for users to make a query via a search engine using their voice. They can either do this using a smartphone or home assistant. This article analyses the uniqueness of voice search and why it should be part of your marketing strategy.

It's been almost two years since Pandora launched its on-demand music streaming service. In that time, the company has done a solid job of fixing some of the issues that cropped up at launch and even adding some features the competition hasn't got to yet (like downloading songs to an Apple Watch for offline playback). Today, Pandora's adding another feature that some of its competitors have: Voice Mode. But, as usual, Pandora believes that the amount of information it has on both the music in its catalog as well as its users will set its voice features apart. For starters, Pandora built Voice Mode internally, from the ground up, something Chief Product Officer Chris Phillips says was key in Voice Mode being a more personal music assistant.

Workday held its European Rising conference last year. One of the key themes from the event was how it is embedding AI into its solutions. Having spoken to Chano Fernandez, Co-President Workday early in the week, we also spoke to Barbry McGann, SVP Product Management at Workday later in the conference. The conversation centred around the main message that Workday delivered at its latest conference, AI. One of its recent product innovations was Skills Cloud.

Any system which performs goal-directed continual learning must not only learn incrementally but process and absorb information incrementally. Such a system also has to understand when its goals have been achieved. In this paper, we consider these issues in the context of question answering. Current state-of-the-art question answering models reason over an entire passage, not incrementally. As we will show, naive approaches to incremental reading, such as restriction to unidirectional language models in the model, perform poorly. We present extensions to the DocQA [2] model to allow incremental reading without loss of accuracy. The model also jointly learns to provide the best answer given the text that is seen so far and predict whether this best-so-far answer is sufficient.

Decision Optimization is now available in the Watson Studio ecosystem with a seamless integration of the CPLEX solvers in the Python runtime environment. Watson Studio now provides everything you need to describe your data, gain insight, and make an optimal decision in the very same ecosystem. Get started right away and learn how to make more intelligent marketing and targeting decisions. Decision Optimization is a subset of data science techniques frequently used for prescriptive analytics. Most documented data science use cases are dedicated to revealing or predicting unknown or future data that is not under your control.

Called IQcast, the feature tells users whether they have a low, medium or high chance of dropping below the target blood glucose range within the next one to four hours. These individual-specific predictions are generated by analyzing data collected through Sugar.IQ app and the Guardian Connect device. The Sugar.IQ app is currently available in the App Store for free download. The FDA-cleared app uses IBM Watson Health's AI and analytics tools to help users see how their glucose levels change during the day, and includes a smart food logging system, motivational insights, a glycemic assistant, a data tracker and a glycemic insights feature. Hypoglycemia -- defined by the American Diabetes Association as a blood glucose level lower than 70 mg/dL -- can lead to symptoms ranging from lightheadedness and lethargy to vision impairment and seizures.

The firm's chief strategy officer on how artificial intelligence is changing the game.

End-to-end neural models have made significant progress in question answering, however recent studies show that these models implicitly assume that the answer and evidence appear close together in a single document. In this work, we propose the Coarse-grain Fine-grain Coattention Network (CFC), a new question answering model that combines information from evidence across multiple documents. The CFC consists of a coarse-grain module that interprets documents with respect to the query then finds a relevant answer, and a fine-grain module which scores each candidate answer by comparing its occurrences across all of the documents with the query. We design these modules using hierarchies of coattention and self-attention, which learn to emphasize different parts of the input. On the Qangaroo WikiHop multi-evidence question answering task, the CFC obtains a new state-of-the-art result of 70.6% on the blind test set, outperforming the previous best by 3% accuracy despite not using pretrained contextual encoders.