AlgorithmWatch take on AccessNow map of regulatory proposals for Artificial Intelligence in Europe


https://algorithmwatch.org/en/accessnow-report-maps-ai-strategies-in-europe/?ref=upflow.co

Makes the incredible valid point, which is still missed by many, that:

"the term [Artificial Intelligence] is ill-suited to circumscribe the challenges we face as societies. AI is too broad and fuzzy a concept to form the basis for discussions on regulatory approaches. It combines aspects that should be kept apart – automated cars, high frequency trading, medical diagnosis, search engines and citizen scoring all include “AI” but it makes little sense to treat them all alike because of this."




Latest Update on Eternime

Latest update on Eternime

http://uk.businessinsider.com/eternime-and-replika-giving-life-to-the-dead-with-new-technology-2018-11?r=US&IR=T

"The beta test has more than 40,000 signups, according to Eternime's website, but is so far only in the hands of around 40 people."

CIMON - Virtual Humans in Space!



"CIMON, the world’s first artificial intelligence-enabled astronaut assistant, has made its debut aboard the International Space Station."

(mind you the astronaut looks more like a plastic Action Man than the bot!)

https://newsroom.ibm.com/CIMON-Gets-Started-as-First-AI-Powered-Astronaut-Assistant?social_post=1902399031&linkId=59736898

Tendar AR Virtual Character

Arckon Chatbot


Interesting looking chatbot called Arckon from Don Patrick which seems to take a development line far closer to what I'm after than most I've seen. He reports using a grammatical parser to understand the input, then rather than focus on Parts of Speech he extracts the implied triples in the sentence, and then uses a triples base and inferencing to derive the response, apparently he "formulates his own sentences dynamically by feeding the relevant facts through a complex grammatical template. The only pre-programmed answers are a few dozen standard manners, like "Thank you" and "You're welcome"."

The following diagram shows how the triples extraction works:



And an example exchange:
















Lots more on his website and blog at: http://artistdetective.com/arckon/


Virtual News Readers



Interesting article in the Guardian yesterday about the Chinese state news agency Xinhua creating a virtual doppelgänger of one of their news readers so he can read and front a highly personalised news service.

A far cry from when I was on Op Raleigh and using the Xinhua HF radio teletype feed to create a daily newspaper!


Production Update

The PDF proofs arrived last week and Maggi and I have until 22 Nov to spot as many typos (few), poor grammar (lots) and poor references (ditto) as we can. Overall layout looking very nice.

Robot reading partners

"Researchers at the University of Wisconsin-Madison have built a robot, named Minnie, to serve as a reading buddy to middle school kids, and Minnie’s new friends grew more excited about books and more attached to the robot over two weeks of reading together." https://ukedchat.com/2018/08/23/robot-reading-partners/

Pepper the robot makes parliamentary "appearance"




https://eandt.theiet.org/content/articles/2018/10/pepper-the-robot-questioned-by-house-of-commons-select-committee-on-ai/

The Institution of Engineering and Technology (latterly the IEE) really ought to know better than this. Reading the article it really sounds like Pepper is a "clever" and "articulate" virtual human, rather than effectively (in this instance) a humanoid style loudspeaker. Pepper was also at the Big Question TV debate I was at, and we had to keep reshooting as even the tech to send the single command to say a line kept failing! There is some nice stuff being done with Pepper, and particularly with some of the other robot-human engagement research at Middlesex University, but presenting it in situations like this, and with coverage like this (as with Sophia and the UN presentation and getting "citizenship") really confuses expectations of what AI and virtual humans and actually really capable right now.

Rant over!

Soul Machines






Some nice video examples of a commercial and very graphic chatbot system - the faces are still a bit uncanny valley though.

Most Common Job in the USA By State





There is a great animated map over at NPR showing the most common job in each US state from 1978 to 2014 and showing how it changes from Farmers and Secretaries to Truck Drivers and Software Developers/Teachers. What is interesting from a Future of Employment discussion is that 1) in the white collar arena the big shift may have taken place (although I doubt it - interesting that customer service agent doesn’t appear) and that 2)  all those Truck Driver jobs could disappear in the next couple of decades.



Beware Fake AI!

An older post but a nice simple analysis we quote in the book of "marketing" AI by Andrew Lea which exactly aligns with our view of how "AI" is being (mis) used nowadays.

https://www.linkedin.com/pulse/beware-fake-ai-andrew-lea.

HAL's Legacy


Nice SETI talk from Dr. David G. Stork about the legacy of the HAL9000, critiquing a video he made 50 years ago when 2001 came out about how feasible was then, and thinking how far things have moved on since.

https://www.youtube.com/watch?v=Pasp2nRaTY8&t=0s&list=PLw6IJozmaWbTLecGQVT69kwKQUzNI7ET-&index=6


Ben Goertzel on Towards Grand Unified AGI


Posting by Ben Goertzel on Towards Grand Unified AGI on the Medium blog.

He details 4 key needs:
  • More effective and fluid interconnection of different AIs into processing networks (which is the core thing the SingularityNET technical design solves).
  • Incentivising of more developers to work on AGI rather than purely on AI that serves the business models of large tech companies and military/intel agencies (again something SingularityNET addresses from an economic/organizational perspective).
  • Rich interoperation of sensory, motoric and cognitive subsystems, so that AIs can learn better from the natural and human world (SingularityNET is working on this via its collaboration with Hanson Robotics).
  • Effective meta-learning — general-purpose learning algorithms that can learn how to learn better, learn how to best carry out learning in particular contexts, learn how to learn and also learn how to learn how to learn.

https://blog.singularitynet.io/toward-grand-unified-agi-c52d75c2f10e

The post focuses mainly on the last point.

Doonesbury 30/9/18



I'm a long time Doonesbury fan so great to see a recent strip moving slightly into virtual humans territory - will be interesting to see if he develops this further!

https://www.gocomics.com/doonesbury/2018/09/30 for the full strip.

Fable Studio's "The Lucy Project"



Whilst in Montreal I got the chance to see Wolves in the Wall, a VR piece by Fables Studio.

A young girl alone in an attic bedroom can hear wolves in the wall - can you help her? This piece, based on a book by Neil Gaiman,  probably had the best realised virtual human I've seen so far in VR. As you follow her around the small room, crouch down to talk to her or look at what she's drawn it really does feel like your in the room with her. The most notable feature was that when she speaks she always looks at you - no matter where you are standing - I kept try to move to catch her out but never did - awesome response to the headset tracking!

It's interesting that Fable, the studio behind this, has kicked off a whole "Lucy Project" to create a character that is interactive and believable. In many ways Lucy is the closest thing I've seen yet to the old Milo "smoke and mirrors" demo from Lionheart - and of course all the more possible because it has digitised the human so everyone is in the same virtual space. You can read more about Wolves in the Walls at https://fable-studio.com/wolves-in-the-walls/ and https://www.oculus.com/blog/vr-visionaries-fable-studio/.

Update:




BQBC Montreal Conference on AI



I was across in Canada last week addressing the BQBC AI Conference in Montreal - a hotbed of Canadian (and global) AI activity. I ended up with two slots, one on the current regulation of AI landscape in the UK (just as well we've been to some of the APPG AI events and submitted evidence to the House of Lords Committee on AI), and may main one on my thoughts on AI - particularly our AI landscape diagram and our work at Daden on virtual tutors, life coaches and personas. There were some great talks from prominent researchers at McGill, and from Microsoft and Arup (on AI and smart cities). That said almost every talk bar mine was on narrow applications of machine learning.


Production Update - September

The revised copy of the book after peer review is now back with the publishers. We've now got the fun task of designing the cover and have a great image in mind. We've been given a target publication date on Feb 2019. Our next big task will probably be indexing the thing!

We've also just changed the theme and layout of the blog, and over the next month will start adding content to each of the chapter pages.

Artificial Intelligence, International Competition, and the Balance of Power


Artificial Intelligence, International Competition, and the Balance of Power

Michael Horowitz evaluates how developments in artificial intelligence — advanced, narrow applications in particular — are poised to influence military power and international politics.


https://tnsr.org/2018/05/artificial-intelligence-international-competition-and-the-balance-of-power/

H2020 FET Flagship Human Brain Project

The Human Brain Project is a H2020 FET Flagship Project which strives to accelerate the fields of neuroscience, computing and brain-related medicine. This acceleration will be achieved by a strategic alignment of scientific research programmes in fundamental neuroscience, advanced simulation and multi-scale modelling with the construction of an enabling Research Infrastructure. https://www.humanbrainproject.eu/en/

Sandra Podcast



A Virtual Human related podcast drama set in a company making an Alexa style smart-speaker, and virtual assistant Sandra.