Large-scale classification, where we need to classify hundreds of thousands or millions of items into thousands of classes, is becoming increasingly common in this age of Big Data… So far, however, very little has been published on how large-scale classification has been carried out in practice, even though there are many interesting questions about such cases. Today's paper is a case study on large-scale classification of products at Walmart. The requirement is to classify 10M products into 5000 categories based on fairly minimal product descriptions. Oh, and new products turn up all the time, and the set of categories is continuously evolving. Many learning solutions assume that we can take a random sample from the universe of items, manually label the sample to create training data, then train a classifier.

For years, machine learning has been used for image, video, and text recognition, as well as serving as the power behind recommendation engines. Today, it's being used to fortify cybersecurity, ensure public safety, and improve medical outcomes. It can also help improve customer service and make automobiles safer. "Machine learning allows you to look at volumes of data and do volumes of calculations that a person really can't do," said Lisa Dolev, founder and CEO of operational intelligence solutions provider Qylur, in an interview. Machine learning can identify patterns that humans tend to overlook or may be unable to find as fast in vast amounts of data.

This paper addresses the problem of detecting information about events from unstructured text. An event-detection system, DISCERN, is presented; its three variants DISCERN- R (rule-based), DISCERN-ML (machine-learned), and DISCERN-C (combined), were evaluated in the NIST TAC KBP 2015 Event Nugget Detection and Event Argument Extraction and Linking tasks. Three contributions of this work are: (a) an approach to collapsing support verb and event nominals that improved recall of argument linking, (b) a new linguist-in-the-loop paradigm that enables quick changes to linguistic rules and examination of their effect on pre- cision and recall at runtime, (c) an analysis of the synergy between the semantic and syntactic features. Results of experimentation with event-detection approaches indicate that linguistically-informed rules can improve precision and machine-learned systems can improve recall. Future refinements to the combination of linguistic and machine learning approaches may involve making better use of the complementarity of these approaches.

TokensRegex is a generic framework included in Stanford CoreNLP for defining patterns over text (sequences of tokens) and mapping it to semantic objects represented as Java objects. TokensRegex emphasizes describing text as a sequence of tokens (words, punctuation marks, etc.), which may have additional attributes, and writing patterns over those tokens, rather than working at the character level, as with standard regular expression packages. TokensRegex was used to develop SUTime, a rule-based temporal tagger for recognizing and normalizing temporal expressions. An included set of slides and the javadoc for TokenSequencePattern provide an overview of this package. Some additional information is available in some older slides.

In today's post, we dive into understanding Association Rules for Market Basket Analysis and discuss three numeric measures that should be considered before deciding to act on / make a business decision based on associations that have been observed in the data: (1) Support (2) Confidence and (3) Lift. Association rules are typically written in the format: Left hand side Implies Right hand side The left hand side is referred to as the Antecedent and the right hand side is the Consequent. The Antecedent means a thing that logically precedes another while a Consequent means a thing that follows as a result. For example, in the association rule: {Butter, Eggs} Implies {Bread} Butter and eggs are the Antecedent while Bread is the Consequent. What this rule means that if you were to pick a shopping cart at random and find butter and eggs in there, there is a chance that you are also likely to find bread.

In warehouses, hospitals and retail stores, and on city streets, industrial parks and the footpaths of college campuses, the first representatives of this new invading force are starting to become apparent. "The robots are among us," says Steve Jurvetson, a Silicon Valley investor and a director at Elon Musk's Tesla and SpaceX companies, which have relied heavily on robotics. A multitude of machines will follow, he says: "A lot of people are going to come in contact with robots in the next two to five years." The arrival of the robots -- and their potentially devastating effect on human employment -- has been widely predicted. Now, the machines are starting to roll or walk out of the labs.

Using humanoid robots to study human behavior.

In the new digital era, a business needs a store of data to help inform its decisions and interactions, but data is useless unless someone acts upon it. Unfortunately, it's very easy to collect more data than any human team could possibly sort through, let alone put into practice. That's where machine learning comes in. These systems learn from information stores to locate patterns and create rules with computer speed. Machine learning is becoming a valuable, perhaps even necessary, part of business infrastructure.

China and Russia said Monday that the South China Sea dispute should not be internationalized and called for its settlement based on negotiation and consultation, Beijing's official Xinhua News reported. The comments come at time when the United States has beefed up its military presence in the contested region in a bid to help the Philippines and other Southeast Asian countries tackle China's assertiveness. Chinese Foreign Minister Wang Yi and his Russian counterpart Sergey Lavrov made the comments during a meeting on Monday. Wang insisted that China was protecting its rights and interests in South China Sea, and was free to choose how to resolve tensions in the area, Xinhua reported. The world's second largest economy's refusal of the Philippines' proposed international arbitration case over the matter was meant to uphold the dignity and authority of the law, Wang said, adding that China and Russia should be cautious against abuses of mandatory arbitration. Meanwhile, Hugo Swire, the British minister of state responsible for East Asia, said earlier in the day that a ruling -- expected in May or early June -- in the Philippines' international arbitration case against China's South China Sea claims must be binding.

The concept of "affordance" represents the relationship between human perceivers and their environment. Affordance perception, representation, and inference are central to commonsense reasoning, tool-use and creative problem-solving in artificial agents. Existing approaches fail to provide flexibility with which to reason about affordances in the open world, where they are influenced by changing context, social norms, historical precedence, and uncertainty. We develop a formal rules-based logical representational format coupled with an uncertainty-processing framework to reason about cognitive affordances in a more general manner than shown in the existing literature. Our framework allows agents to make deductive and abductive inferences about functional and social affordances, collectively and dynamically, thereby allowing the agent to adapt to changing conditions. We demonstrate our approach with an example, and show that an agent can successfully reason through situations that involve a tight interplay between various social and functional norms.