Semantic text matching is a critical problem in information retrieval. Recently, deep learning techniques have been widely used in this area and obtained significant performance improvements. However, most models are black boxes and it is hard to understand what happened in the matching process, due to the poor interpretability of deep learning. This paper aims at tackling this problem. The key idea is to test whether existing deep text matching methods satisfy some fundamental heuristics in information retrieval. Specifically, four heuristics are used in our study, i.e., term frequency constraint, term discrimination constraint, length normalization constraints, and TF-length constraint. Since deep matching models usually contain many parameters, it is difficult to conduct a theoretical study for these complicated functions. In this paper, We propose an empirical testing method. Specifically, We first construct some queries and documents to make them satisfy the assumption in a constraint, and then test to which extend a deep text matching model trained on the original dataset satisfies the corresponding constraint. Besides, a famous attribution based interpretation method, namely integrated gradient, is adopted to conduct detailed analysis and guide for feasible improvement. Experimental results on LETOR 4.0 and MS Marco show that all the investigated deep text matching methods, both representation and interaction based methods, satisfy the above constraints with high probabilities in statistics. We further extend these constraints to the semantic settings, which are shown to be better satisfied for all the deep text matching models. These empirical findings give clear understandings on why deep text matching models usually perform well in information retrieval. We believe the proposed evaluation methodology will be useful for testing future deep text matching models.

If you are a fan of Iron Man or the Avengers franchise, you know that when Tony Stark is talking to Jarvis, he is talking with Artificial Intelligence or AI. According to computer science, an'intelligent machine' is a rational agent who perceives its environment and takes those actions, which will maximize its chances of success for the specified goal. In simple terms, intelligence that is exhibited by the machines is termed as'Artificial Intelligence' or AI. Ever wondered how Google Now, Siri, and Cortana help you find information? They are all AI applications used on different platforms.

If you are a fan of Iron Man or the Avengers franchise, you know that when Tony Stark is talking to Jarvis, he is talking with Artificial Intelligence or AI. According to computer science, an'intelligent machine' is a rational agent who perceives its environment and takes those actions, which will maximize its chances of success for the specified goal. In simple terms, intelligence that is exhibited by the machines is termed as'Artificial Intelligence' or AI. Ever wondered how Google Now, Siri, and Cortana help you find information? They are all AI applications used on different platforms.

Facebook's new Deep Text system, launched last month, promises to revolutionise the social media experience for both individual users and brands. According to the company, its artificial intelligence (AI) engine can analyse and understand the meaning and sentiment in the text of thousands of posts per second across 20 languages. It is said to be able to achieve "near human accuracy" in understanding the deeper meaning behind written words and distinguishing between slang, colloquialisms and brand names. Facebook has already used the technology in its Messenger app to help users hail taxis and on its main site to advise users who are making sales. But Deep Text could also have a huge impact on the way brands target and communicate with their customers.

"AI is the new mobile" is a phrase we might as well start getting used to hearing. It's infuriating that, just as the marketing world finally starts to think "mobile first", not one but two technology giants signal loudly that they're over mobile and on to the next thing. Witness Google chief executive Sundar Pichai, who recently spoke about moving from a mobilefirst to an AI-first world. Make no mistake, he sees artificial intelligence as the future of search, best exemplified by what Google calls the Google assistant: less a product and more an artificially intelligent, conversational interface to all things Google. Earlier in the year, at Facebook's developer conference F8, Mark Zuckerberg made AI one of the three pillars of Facebook's ten-year roadmap (alongside – yes, you guessed it – connectivity and virtual reality).

Facebook is rolling out a "deep-learning based text understanding engine" that leverages neural network architectures and promises to challenge the Google search engine by bringing "near-human accuracy" to the process of sorting through the daily flood of social media posts on its web site. Facebook's (NASDAQ: FB) AI initiative called Deep Text announced in a blog post last week also could advance the state of the art in terms of organizing the flood of unstructured text data generated by the leading social media site. The approach also seeks to expand current natural language processing (NLP) techniques that are frequently tripped up by slang. Among the early goals of the search effort is ferretting out relevant information from social media posts that would link sellers with buyers. The company said Deep Text uses convolutional and recurrent neural network approaches to perform word- and character-level learning. The scheme also employs Torch to train neural networks along with its "AI backbone called Fb Learner Flow.

Facebook this week pulled the curtains off Deep Text, its new deep-learning-based text-understanding engine, which it said can analyze the textual content in several-thousand posts per second. The social network said in a blog post by engineer Ahmad Abdulkader, applied machine learning platform team technical program manager Aparna Lakshmiratan and research scientist Joy Zhang that the aim of Deep Text is to better understand text across Facebook and to reduce reliance on language-dependent knowledge, adding that the technology is already being tested on Messenger and offering the example that Deep Text can decide whether a user is looking for a taxi by distinguishing between "I just came out of the taxi" and "I need a ride." Text understanding includes multiple tasks, such as general classification to determine what a post is about--basketball, for example--and recognition of entities, like the names of players, stats from a game and other meaningful information. But to get closer to how humans understand text, we need to teach the computer to understand things like slang and word-sense disambiguation. As an example, if someone says, "I like blackberry," does that mean the fruit or the device?

He said advanced AI robotics will eliminate jobs, and that there is also a real issue with making sure humans retain control of machines at all times. Echoing Gates' warning over AI, Tesla's CEO Elon Musk also shared his worry that AI is only used by a small percentage of people "to the detriment of humanity as a whole". Although excited about AI developments, Musk told the Live Code audience that "not all AI futures are benign,' and that very real dangers are associated with the technology. The other two big headliners at the conference were Google CEO Sundar Pichai and Facebook director Hussein Mehanna, with both talking strategy and their companies objectives in the AI market. Google's Pichai confidently stated that his company was in the AI game to win it. He said: "For us, we definitely see a huge opportunity.

While Facebook has lot of noise around using artificial intelligence to better sort photos and videos, text is still a huge part of the Facebook experience. Today, we got a look behind the scenes on how AI is helping Facebook sift all that text and improve the Facebook experience with "Deep Text" -- a system developed by Facebook's AI labs that scans 10,000 posts every second in 20 languages. People post more than 1 billion items -- statuses, links, photos, whatever -- to Facebook every day, says a company spokesperson. Deep Text is "deep learning-based text understanding engine," as Facebook puts it. Its ability to understand text like a human would is already being put to work in Facebook Messenger.

While Facebook has lot of noise around using artificial intelligence to better sort photos and videos, text is still an huge part of the Facebook experience. Today, we got a look behind the scenes on how AI is helping Facebook sift all that text and improve the Facebook experience with "Deep Text" -- a system developed by Facebook's AI labs that scans 10,000 posts every second in 20 languages. People post more than 1 billion items -- statuses, links, photos, whatever -- to Facebook every day, says a company spokesperson. Deep Text is "deep learning-based text understanding engine," as Facebook puts it. Its ability to understand text like a human would is already being put to work in Facebook Messenger. And Facebook plans to take it a lot further.