How Alexa Is Learning to Converse More Naturally : Alexa Blogs

#artificialintelligence

To handle more-natural spoken interactions, Alexa must track references through several rounds of conversation. If, for instance, a customer says, "How far is it to Redmond?" and after the answer follows up by saying, "Find good Indian restaurants there", Alexa should be able to infer that "there" refers to Redmond. We call the task of reference tracking "context carryover," and it's a capability that is currently being phased in to the Alexa experience. At this year's Interspeech, the largest conference on spoken-language understanding, my colleagues and I will present a paper titled "Contextual Slot Carryover for Disparate Schemas," which describes our solution to the problem of slot carryover, a crucial aspect of context carryover. "Domain" describes the type of application -- or "skill" -- that the utterance should invoke; for instance, mapping skills should answer questions about geographic distance.


Developing a business strategy by combining machine learning with sensitivity analysis Amazon Web Services

#artificialintelligence

Machine learning (ML) is routinely used by countless businesses to assist with decision making. In most cases, however, the predictions and business decisions made by ML systems still require the intuition of human users to make judgment calls. In this post, I show how to combine ML with sensitivity analysis to develop a data-driven business strategy. This post focuses on customer churn (that is, the defection of customers to competitors), while covering problems that often arise when using ML-based analysis. These problems include difficulties with handling incomplete and unbalanced data, deriving strategic options, and quantitatively evaluating the potential impact of those options.


Developing a business strategy by combining machine learning with sensitivity analysis Amazon Web Services

#artificialintelligence

Machine learning (ML) is routinely used by countless businesses to assist with decision making. In most cases, however, the predictions and business decisions made by ML systems still require the intuition of human users to make judgment calls. In this post, I show how to combine ML with sensitivity analysis to develop a data-driven business strategy. This post focuses on customer churn (that is, the defection of customers to competitors), while covering problems that often arise when using ML-based analysis. These problems include difficulties with handling incomplete and unbalanced data, deriving strategic options, and quantitatively evaluating the potential impact of those options.


Two New Papers Discuss How Alexa Recognizes Sounds : Alexa Blogs

#artificialintelligence

Last year, Amazon announced the beta release of Alexa Guard, a new service that lets customers who are leaving the house instruct their Echo devices to listen for glass breaking or smoke and carbon dioxide alarms going off. At this year's International Conference on Acoustics, Speech, and Signal Processing, our team is presenting several papers on sound detection. I wrote about one of them a few weeks ago, a new method for doing machine learning with unbalanced data sets. Today I'll briefly discuss two others, both of which, like the first, describe machine learning systems. One paper addresses the problem of media detection, or recognizing when the speech captured by a digital-assistant device comes from a TV or radio rather than a human speaker.


Re-Ranking Recommendations Based on Predicted Short-Term Interests - A Protocol and First Experiment

AAAI Conferences

The recommendation of additional shopping items that are potentially interesting for the customer has become a standard feature of modern online stores. In academia, research on recommender systems (RS) is mostly centered around approaches that rely on explicit item ratings and long-term user profiles. In practical environments, however, such rating information is often very sparse and for a large fraction of the users very little is known about their preferences. Furthermore, in particular when the shop offers products from a variety of categories, the decision of what should be recommended can strongly depend on the user's current short-term interests and the navigational context. In this paper, we report the results of an initial experimental analysis evaluating the predictive accuracy of different contextualized and non-contextualized recommendation strategies and discuss the question of appropriate experimental designs for such types of evaluations. To that purpose, we introduce a parameterizable protocol that supports session-specific accuracy measurements. Our analysis, which was based on log data obtained from a large online retailer for clothing and lifestyle products, shows that even a comparably simple contextual post-processing approach based on product features can leverage short-term user interests to increase the accuracy of the recommendations.