Theoretical Analysis of Heuristic Search Methods for Online POMDPs

Neural Information Processing Systems

Planning in partially observable environments remains a challenging problem, despite significant recent advances in offline approximation techniques. A few online methods have also been proposed recently, and proven to be remarkably scalable, but without the theoretical guarantees of their offline counterparts. Thus it seems natural to try to unify offline and online techniques, preserving the theoretical properties of the former, and exploiting the scalability of the latter. In this paper, we provide theoretical guarantees on an anytime algorithm for POMDPs which aims to reduce the error made by approximate offline value iteration algorithms through the use of an efficient online searching procedure. The algorithm uses search heuristics based on an error analysis of lookahead search, to guide the online search towards reachable beliefs with the most potential to reduce error.


Google updates Maps, Search and Assistant so you can order food without app

USATODAY - Tech Top Stories

Google has unveiled updates for its artificially intelligent voice assistant and new privacy tools to give people more control over how they're being tracked on the go or at home. The company also unveiled a new Pixel phone and smart home display. Google just made ordering pizza, pad thai and fried chicken from your favorite restaurants even easier. The search giant announced on Thursday that it updated apps like Google Maps, Google Search and the Google Assistant to make ordering food online more convenient, so you don't have to download as many third-party apps. "When I was pregnant with my son last year, my cravings were completely overpowering," said Google's senior product manager of food ordering, Anantica Singh, in a blog post.


Semantic Indexing: Google's Big Data Trick For Multilingual Search Results

#artificialintelligence

Google has perfected its ability to execute web search results for its users all over the world. In the early days of the Internet, the search engine was primarily suited for displaying search results for English users. Non-English-speaking users have complained that search results are often displayed in the wrong language entirely. However, Google is becoming more proficient at providing search results in other languages as well. A lot of factors can play a role, but one of the biggest is its use of deep learning to understand semantic references--enter semantic indexing.


Why Machine Learning Is Key to the Search Marketing of Tomorrow

@machinelearnbot

Advertising has changed a lot over the years. There was a time when machine learning, automation, and software-based marketing tech stacks weren't a "thing." There are hundreds of channels across physical and print media and online at present, including social, mobile, and video. Even TV has diversified into hundreds of cable channels on your remote control. And yet, digital ad revenue has gone on to surpass that of TV.


AI For eMerchants: Beyond Visual Search Engines PYMNTS.com

@machinelearnbot

When is milk not milk? This is no trick question -- it's a distinction that artificial intelligence (AI) is going to have to learn to make in order for eMerchants to fully leverage the potential of machine learning. To one customer, "buy milk" means buy a gallon of whole milk; to another, a 1.4-liter jug of unsweetened vanilla almond milk. Digital shopping lists, apps and virtual assistants must understand this and not force the customer to spell it out each time before these platforms can successfully become the new normal. "We think about things in shorthand, not in terms of specifics," Dave Barrowman, Skava VP of Innovation, told PYMNTS' Karen Webster in a recent webinar.