Mensing.AI


Learnings & Musings on AI, ML, Data Science & Python

What’s That Animal? ๐Ÿ˜๐Ÿฆ๐Ÿƒ

Snapshot Serengeti is investigating the use of deep learning image recognition to sift through their trove of 3.2 million wildlife pictures from various camera traps. This is an AI researcher/engineer’s dream data set, there is just so much. ๐Ÿ“Š So what can this model do? ๐Ÿ” Not only does the artificial intelligence system tell you which of 48 different species … Read More


Alexa, Update ๐Ÿ–ฒ

Tech Fest season continues! Amazon announced some cool updates to Alexa via blog and conference appearance. ๐ŸŽŠ These updates will make interacting with Alexa feel more like interacting with a human. Not because she (it?) is getting and smarter necessarily, but because the interface is improving. Amazon is continuing to put the customer experience at the forefront, which is why … Read More


Generate an Adversary ๐ŸฅŠ

This is a really great introduction to Generative Adversarial Networks (GANs) with great visuals. (and a tutorial if you want to get your hands dirty) ๐Ÿ‘ฉโ€๐Ÿ’ป๐Ÿ‘จโ€๐Ÿ’ป Why should you care about GANs? They are the driving force behind the crazy creative abilities of neural networks. They are also the building blocks of deepfakes in many cases. ๐Ÿ– What’s the 101 … Read More


The Revolution Hasnโ€™t Happened Yet. โœŠ No Word on if it will be Televised. ๐Ÿ“บ

Michael does a great job breaking down the difference between the meaning of AI and what is currently being billed as AI (mostly ML). And setting the stage for what could be if AI became one avenue to explore. ๐Ÿ”ฎ let us conceive broadly of a discipline of โ€œIntelligent Infrastructureโ€ (II), whereby a web of computation, data and physical entities … Read More


AI Coming to the Mobile Masses ๐Ÿ“ฒ

Applepalooza 2018 is happening now! (a.k.a. WWDC) This year’s focus is software from what I hear, and they went right for the trendy trend by announcing a new AI/ML platform that will allow easy rollout for developers in the App Store. ๐Ÿ Called Create ML, it will allow devs to easily add machine learning capabilities to their apps using Apple’s … Read More


This Computer is Super ๐Ÿ’ป๐Ÿ’ช

One supercomputer to rule them all! (For now) Oak Ridge Lab in the US has unveiled it’s new supercomputer, Summit, and takes the crown for most powerful computer in the world, claiming the title from China. ๐Ÿ‘‘๐Ÿ‡บ๐Ÿ‡ธ How powerful is it? everyone on earth would have to do a calculation every second of every day for 305 days to crunch … Read More


Bees are Zen, Care about Nothing ๐Ÿ

Bees have officially reached stage 3 of 4 in understanding the concept of zero. ๐ŸŽ‰ Beginning to understand how bees learn and understand things, and potential transfer learning throughout the hive, could be really interesting for AI. I’ve been curious why we’re so focused on recreating human thought in machines, especially when we still seem to understand so little about … Read More


Accenture Goes to Ethicsville

Accenture has entered the bias-beating ring, they’ll be formally announcing their fairness tool this week. I’m glad to see a growing surge of focus on algorithmic bias and how to avoid it. Are any of these tools silver bullets? No. But acknowledging the problem and seeking to solve even small portions of it is a wonderful first step. ๐Ÿ—บ It’s … Read More


Whatโ€™s Your Rorschach?

My general line of thinking towards the fear mongering around Artificial General Intelligence was that it was ratcheting up fear of the unknown at the expense of much more pressing things. Like algorithmic bias, weaponized advertising and social media, and humans in general. Open a newspaper, do we really need to worry about the potential future of Terminator becoming real … Read More


New from Google: AI Ethics โš–๏ธ

Google has released a new set of ethics for AI applications, likely in response to the PR disaster that was their partnership with the DoD (and seriously, is that patch for real?). But that doesn’t mean their new principles are bad, they look good to me. Be socially beneficial Avoid creating or reinforcing unfair bias Be built and tested for … Read More