Daily Archives: June 17, 2017

TechCrunch

A discussion about AI's conflicts and challenges TechCrunch He's now research director at the Alan Turing Institute in the UK, which aims to support data science research, which of course means machine learning and AI, and includes probing the ethics and societal implications of AI and big data. So how can a ...

Here is the original post:

A discussion about AI's conflicts and challenges - TechCrunch

There appears to be a rather strange correlation between artificial intelligence and gaming right now. We have seen AIs take on players in Go, and Microsoft is now tackling the infamous Ms. Pac-Man game. Their new AI has claimed the top score in this particular game, which makes it four times better than the top-ranking human player. All of this shows us humans are not necessarily the greatest at video games, unfortunately.

It is not entirely surprising to learn artificial intelligence solutions are better at playing particular video games compared to most humans. After all, we humans are great at doing several things quite well, whereas an AI can excel at doing just one thing better than anything else. In this particular case, Microsofts new AI has thrashed all human players when it comes to Ms. Pac-Man, as it successfully reached the top score possible. An impressive feat, although some people may not be too happy about it.

To put this news into perspective, reaching the top score in Ms. Pac-Man is no easy feat, not even for an AI. In fact, experts felt his particular game is the hardest to beat for an artificial intelligence. There is a lot going on at the same time, and the circumstances of the game can shift around quite a bit. Interestingly enough, Microsofts concoction was more than capable of handling the pressure and made playing the game look like a walk in the park.

The AI in question is developed by Maluuba, a Canadian deep learning startup owned by Microsoft. The company was acquired earlier in 2017, and it appears Microsoft has made a smart move in doing so. Beating the Ms. Pac-Man game is the companys first major feat, but it certainly is one that will go down in the history books. In comparison, the best human player to date scored four times fewer points in the game.

Under the hood, the AI used a method described as Hybrid Reward Architecture. It is a combination of reinforcement learning algorithms. By training around 150 individual AI agents to work together in parallel, the team successfully ensured its creation could reach the top score. All agents were rewarded for eating pellets, whereas others were trained to avoid ghosts. Combining all of these experiences into one project lead to this unexpected success.

The world of video games is slowly being dominated by artificial intelligence these days. Googles AlphaGothrashed the world champion not too long ago and promptly announced retirement. Now that Microsofts project has beaten Ms. Pac-Man, it is doubtful the AI will play this game ever again. Once the maximum result is achieved, there is no reason for these AI creations to keep playing the same game, as there is no further room for significant improvement.

The fact an AI can defeat a rather unpredictable game with such ease is rather telling, though. It is evident artificial intelligence is incredibly powerful at this stage, at least where these single-player games are concerned. It will take some time until we see an AI be successful in competitive multiplayer games, though, but it may happen at some point.

If you liked this article, follow us on Twitter @themerklenews and make sure to subscribe to our newsletter to receive the latest bitcoin, cryptocurrency, and technology news.

See more here:

Microsoft's New AI Reaches High Score in Ms. Pac-Man - The Merkle

Intel and Microsoft have been on something of an artificial intelligence (AI) investment binge of late, with the chip and software giants announcing a slew of deals this week via their respective VC arms Intel Capital and Microsoft Ventures.

Perhaps the most notable of these was Element AI, which raised a gargantuan $102 million in what is one of the largest series A rounds in recent times. The Montreal-based startup, which helpsconnect companies with machine learning experts, drew in some other interesting investors besides Intel and Microsoft, including rival chipmaker Nvidia.

The Element AI deal followed just a day afterIntel and Microsoft joined forcesfor a $15 million investment into CognitiveScale, a Texas-based startup that uses AI to harness big data and deliver insights and recommendations. The very same day, Intel participated in a $16 million round into California-based robotic vision startup Aeye, while on Monday Microsoft got involved in a $20 million funding round into CrowdFlower, a platform that meshes machines with human input to ensure data science teams have access to properly tagged, clean data. Microsoft also invested in in CrowdFlower last year.

The duo have been investing in AI-related startups all year, too.

Back in January, Intel Capitaljoined a $14 million funding round into Mighty AI, astartup that uses the general public to help train companies AI algorithms, and a $7 million round into cybersecurity startupFortscale, which uses machine learning and big data analytics to detect malicious user behavior. But the most sizeable deal to emerge from Intel this year so far was its $15 billion acquisition of Mobileye, a computer vision firm specializing in autonomous cars.

Last month, Microsoft Ventures revealed it had contributed to a $3.5 million seed round intoNew York-based Agolo, a startup thats setting out to help companiesfight information overload through AI-powered summarizations. And on the same day, Microsoft Ventures announced it was also leading a $7.6 million round into Bonsai, a startup that helps companies build AI into their business.

As with Intel, Microsoft has also made some AI acquisitions this year. Last week, itacquired Israeli startup Hexadite to bring AI-powered security to Windows 10 in the enterprise, while back in January it boughtresearch-oriented deep learning startup Maluuba.

All the major tech titans, from Alphabet and Amazon through to Apple and beyond, have made AI acquisitions in recent times, but the recent flurry over the past week indicates the talent land grab is intensifying. Indeed, Microsoft recently launched a dedicated fund for AI startups, while Intel Capital alone now claims more than 25 AI companies in its portfolio.

As we proceed toward realizing the promise and opportunity offered by AI, Intel is aggressively investing in the best minds in the industry that bring the skills and technology necessary to accelerate complex, data-intensive processes, Intel said in a statement issued to VentureBeat. Each of these companies is involved in furthering innovations with AI ranging from robotics and autonomous vehicles to machine intelligence software and an AI research lab with the premise of solving our most daunting challenges by accelerating large-scale problem-solving, unleashing new scientific discovery, extending our human senses and capabilities, and relieving us of the burden of tedious or dangerous tasks.

See the original post:

Intel and Microsoft's latest investment binge shows AI land grab is intensifying - VentureBeat

In Brief While developing negotiating chatbot agents, Facebook researchers found that the bots spontaneously developed their own non-human language as they improved their techniques, highlighting how little we still know about how artificial intelligences learn. The Future of Language

A recent Facebook report on the way chatbots converse with each other has given the world a glimpse intothe future of language.

In the report, researchers from the Facebook Artificial Intelligence Research lab (FAIR) describe training their chatbot dialog agents to negotiate using machine learning. The chatbots were eager and successful dealmaking pupils, but the researchers eventually realized they needed to tweak their model because the bots were creating their own negotiation language, diverting from human languages.

To put it another way, when they used a model that allowed the chatbots to converse freely, using machine learning to incrementally improve their conversational negotiation strategies as they chatted, the bots eventually created and used their own non-human language.

The unique, spontaneous development of a non-human language was probably the most baffling and thrilling development for the researchers, but it wasnt the only one. The chatbots also proved to be smart about negotiating and used advanced strategies to improve their outcomes. For example, a bot might pretend to be interested in something that had no value to it in order to be able to sacrifice that thing later as part of a compromise.

Although Facebooks bargain-hunting bots arent a sign of an imminent singularity or anything even approaching that level of sophistication they are significant, in part because they prove once again that an important realm we once assumed was solely the domain of humans, language, is definitely a shared space. This discovery also highlights how much we still dont know about the ways that artificial intelligences (AIs) think and learn, even when we create them and model them after ourselves.

Read more:

A Facebook AI Unexpectedly Created Its Own Unique Language - Futurism