recentpopularlog in

pierredv : ai   131

« earlier  
AI-powered smartphone cameras are changing the way we see reality | New Scientist, issue 3221, Mar 2019
"Smartphone cameras now use artificial intelligence to completely transform the pictures we take, and it could change the way we see reality"

"The goal of digital photography was once to approximate what our eyes see. “All digital cameras, including ones on smartphones, have always had some sort of processing to modify contrast and colour balance,” says Neel Joshi, who works on computer vision at Microsoft Research.
Computational photography goes beyond this, automatically making skin smoother, colours richer and pictures less grainy. It can even turn night into day. These photos may look better, but they raise concerns about authenticity and trust in an era of fakeable information. “The photos of the future will not be recorded, they’ll be computed,” says Ramesh Raskar at the MIT Media Lab. "

"Other clever things your phone’s camera can do

Google Translate’s camera feature lets you point your phone at a sign in a foreign country or words in a book (see picture), superimposing an on-screen translation of the text to make it appear in your own language.

Microsoft’s Pix app detects if you are taking a picture of a whiteboard or receipt, automatically straightening and cropping the image to the text contained within. The firm’s Excel app can capture printed data tables and automatically put them into a spreadsheet, doing away with the need for tedious data entry.

Samsung’s Bixby virtual assistant has a calorie counter built into its camera (see picture below). Point it at what is on your plate and it will give you an estimate of the food’s energy content.
"
photography  AI  cameras  NewScientist 
7 days ago by pierredv
FACT CHECK: Did Facebook Shut Down an AI Experiment Because Chatbots Developed Their Own Language?, Aug 2017
"Facebook's artificial intelligence scientists were purportedly dismayed when the bots they created began conversing in their own private language."

Rated FALSE
Snopes  Facebook  AI  language  bots 
21 days ago by pierredv
How AI Will Help Radar Detect Tiny Drones 3 Kilometers Away - Defense One
via Brian Daly, AT&T

"AESA radars, which steer its multiple radar beams electronically instead of using physical gimbals, have been around for years. The real innovation lies in training software to detect objects, including objects as small as DJI’s popular Mavic drones, in radar imagery. But there’s very little imagery data to train a machine learning algorithm on how to see something that small. What’s needed is a dataset of extremely small modulations in the echoes of radar signals. The researchers used a GAN to turn a small bit of available training data into an abundance."
AI  radar  drones  UAS 
24 days ago by pierredv
Algorithmic Transparency for the Smart City by Robert Brauneis, Ellen P. Goodman :: SSRN
Via Blake Reid, Jul 2019
"To do this work, we identified what meaningful “algorithmic transparency” entails. We found that in almost every case, it wasn’t provided. Over-broad assertions of trade secrecy were a problem. But contrary to conventional wisdom, they were not the biggest obstacle. It will not usually be necessary to release the code used to execute predictive models in order to dramatically increase transparency. We conclude that publicly-deployed algorithms will be sufficiently transparent only if (1) governments generate appropriate records about their objectives for algorithmic processes and subsequent implementation and validation; (2) government contractors reveal to the public agency sufficient information about how they developed the algorithm; and (3) public agencies and courts treat trade secrecy claims as the limited exception to public disclosure that the law requires. Although it would require a multi-stakeholder process to develop best practices for record generation and disclosure, we present what we believe are eight principal types of information that such records should ideally contain. "
algorithms  AI  governance  smart-cities  SSRN 
4 weeks ago by pierredv
(56) (PDF) Notes on bias in the socio-material realization of AI technologies | Hans Radder - Academia.edu
These notes are a revised version of an 'extended abstract' submitted to, and presented at, the conference on Bias in AI and Neuroscience, 17-19 June 2019, Radboud University Nijmegen.
AI  philosophy  bias  Academia.edu 
6 weeks ago by pierredv
On the dark history of intelligence as domination | Aeon Essays, Stephen Cave , Feb 2017
Intelligence has always been used as fig-leaf to justify domination and destruction. No wonder we fear super-smart robots
AeonMagazine  intelligence  AI 
10 weeks ago by pierredv
An AI conference warns us why we need to mind our language | New Scientist issue 3212, Jan 2019
"We’re using the wrong words to talk about artificial intelligence."

"Language is at the heart of the problem. In his 2007 book, The Emotion Machine, computer scientist Marvin Minsky deplored (although even he couldn’t altogether avoid) the use of “suitcase words”: his phrase for words conveying specialist technical detail through simple metaphors. Think what we are doing when we say metal alloys “remember” their shape, or that a search engine offers “intelligent” answers to a query."

"Without metaphors and the human tendency to personify, we would never be able to converse, let alone explore technical subjects, but the price we pay for communication is a credulity when it comes to modelling how the world actually works. No wonder we are outraged when AI doesn’t behave intelligently. But it isn’t the program playing us false, rather the name we gave it."

"Earlier this year in a public forum [Turkish-born Memo Akten, based at Somerset House in London] threatened to strangle a kitten whenever anyone in the audience personified AI, by talking about “the AI”, for instance."
NewScientist  language  quotes  metaphor  thinking  cognition  AI  anthropomorphism  culture 
11 weeks ago by pierredv
AI narratives: portrayals and perceptions of artificial intelligence and why they matter | Royal Society
In a series of four workshops (PDF), the Royal Society and Leverhulme Centre for the Future of Intelligence explored:

which narratives around intelligent machines are most prevalent, and their historical roots;
what can be learned from how the narrative around other complex, new technologies developed, and the impact of these;
how narratives are shaping the development of AI, and the role of arts and media in this process; and
the implications of current AI narratives for researchers and communicators.
AI  narratives  stories  RoyalSociety 
12 weeks ago by pierredv
Ideas in Excel - Office Support
"Ideas in Excel empowers you to understand your data through high-level visual summaries, trends, and patterns. Simply click a cell in a data range, and then click the Ideas button on the Home tab. Ideas in Excel will analyze your data, and return interesting visuals about it in a task pane."
Microsoft  Office  Excel  howto  AI 
12 weeks ago by pierredv
Botnik Studios - Harry Potter
Via New Scientist, 5 Jan 2019, issue 3211

Apparently uses a predictive text keyboard
AI  prediction  fiction  HarryPotter  spoofing 
may 2019 by pierredv
This wild, AI-generated film is the next step in “whole-movie puppetry” | Ars Technica Jun 2018
Via New Scientist, 5 Jan 2019, issue 3211

"Director Oscar Sharp and AI researcher Ross Goodwin, have returned with another AI-driven experiment that, on its face, looks decidedly worse. Blurry faces, computer-generated dialogue, and awkward scene changes fill out this year's Zone Out, a film created as an entry in the Sci-Fi-London 48-Hour Challenge—meaning, just like last time, it had to be produced in 48 hours and adhere to certain specific prompts.

That 48-hour limit is worth minding, because Sharp and Goodwin went one bigger this time: they let their AI system, which they call Benjamin, handle the film's entire production pipeline."
ArsTechnica  movies  AI 
may 2019 by pierredv
The first piece of AI-generated art to come to auction | Christie's
"The painting, if that is the right term, is one of a group of portraits of the fictional Belamy family created by Obvious, a Paris-based collective consisting of Hugo Caselles-Dupré, Pierre Fautrel and Gauthier Vernier. They are engaged in exploring the interface between art and artificial intelligence, and their method goes by the acronym GAN, which stands for ‘generative adversarial network’."

Via New Scientist, 5 Jan 2019, issue 3211

‘On one side is the Generator, on the other the Discriminator. We fed the system with a data set of 15,000 portraits painted between the 14th century to the 20th. The Generator makes a new image based on the set, then the Discriminator tries to spot the difference between a human-made image and one created by the Generator. The aim is to fool the Discriminator into thinking that the new images are real-life portraits. Then we have a result.’
Christies  art  portraits  AI 
may 2019 by pierredv
xkcd: Machine Learning
From https://www.explainxkcd.com/wiki/index.php/1838:_Machine_Learning

"Cueball stands next to what looks like a pile of garbage (or compost), with a Cueball-like friend standing atop it. The pile has a funnel (labelled "data") at one end and a box labelled "answers" at the other. Here and there mathematical matrices stick out of the pile. As the friend explains to the incredulous Cueball, data enters through the funnel, undergoes an incomprehensible process of linear algebra, and comes out as answers. The friend appears to be a functional part of this system himself, as he stands atop the pile stirring it with a paddle. His machine learning system is probably very inefficient, as he is integral to both the mechanical part (repeated stirring) and the learning part (making the answers look "right"). "
xkcd  comic  cartoons  AI  MachineLearning 
may 2019 by pierredv
In Praise of Artificial Stupidity – Towards Data Science
Via Martin Geddes

"In what follows, I submit my deeply subjective overview of A.I. — yesterday, today, tomorrow — and shall commit the ultimate (nerd) sin: decoupling the scienc-y hype from the business value."

"Funny enough, the difference between humans and machines has become evident not so much in how many answers they got right/wrong, but more in how they are wrong. In other words, even if humans and machine agree on the caption for this picture: ...

they would agree based on very different “thought processes”. As a proof of that, the following machine-generated caption (“a dinosaur on top of a surfboard”) is not just wrong: it’s actually so far from an “intelligent guess” that raises suspicion on the entire game— how is it possible for a physical system that understands images to be so wrong?"

"explainability: it’s often noted that deep learning models are “black boxes”, as it is hard for humans to understand why they do what they do. While most observers stress obvious ethical and practical consequences of such feature, our little tour through A.I. history (doesn’t certainly prove but) suggests that some degree of explainability may be a key part of “intelligence”: having representations implies some sort of “modular” structure which leads more naturally to answer “why questions” than a simple matrix of weights;"
AI  quotations 
may 2019 by pierredv
How Tech Utopia Fostered Tyranny - The New Atlantis -Winter 2019
"Authoritarians’ love for digital technology is no fluke — it’s a product of Silicon Valley’s “smart” paternalism"

"ools based on the premise that access to information will only enlighten us and social connectivity will only make us more humane have instead fanned conspiracy theories, information bubbles, and social fracture. A tech movement spurred by visions of libertarian empowerment and progressive uplift has instead fanned a global resurgence of populism and authoritarianism."

"But what we are searching for — what we desire — is often shaped by what we are exposed to and what we believe others desire. And so predicting what is useful, however value-neutral this may sound, can shade into deciding what is useful, both to individual users and to groups, and thereby shaping what kinds of people we become, for both better and worse."

"As long as our desires are unsettled and malleable — as long as we are human — the engineering choices of Google and the rest must be as much acts of persuasion as of prediction."

"Each company was founded on a variation of the premise that providing more people with more information and better tools, and helping them connect with each other, would help them lead better, freer, richer lives."

"Moreover, because algorithms are subject to strategic manipulation and because they are attempting to provide results unique to you, the choices shaping these powerful defaults are necessarily hidden away by platforms demanding you simply trust them"

"We can see the shift from “access to tools” to algorithmic utopianism in the unheralded, inexorable replacement of the “page” by the “feed.” "

"By consuming what the algorithm says I want, I trust the algorithm to make me ever more who it thinks I already am."

"What’s shocking isn’t that technological development is a two-edged sword. It’s that the power of these technologies is paired with a stunning apathy among their creators about who might use them and how. Google employees have recently declared that helping the Pentagon with a military AI program is a bridge too far, convincing the company to cancel a $10 billion contract. But at the same time, Google, Apple, and Microsoft, committed to the ideals of open-source software and collaboration toward technological progress, have published machine-learning tools for anyone to use, including agents provocateur and revenge pornographers."

"They and their successors, based on optimistic assumptions about human nature, built machines to maximize those naturally good human desires. But, to use a line from Bruno Latour, “technology is society made durable.” That is, to extend Latour’s point, technology stabilizes in concrete form what societies already find desirable."
politics  surveillance  technology  TheNewAtlantis  Google  Facebook  AI  prediction  ethics  morality  search  trust  behavior 
april 2019 by pierredv
The Race Is On: Assessing the US-China Artificial Intelligence Competition - Modern War Institute Apr 2019
A revanchist Russia might be the scourge of the Western defense community, but Vladimir Putin has arguably issued the clearest articulation of AI’s massive potential: “Whoever becomes the leader in [AI] will become the ruler of the world.” But how do we assess who is leading?
AI  US  China  warfare 
april 2019 by pierredv
Top-10 Artificial Intelligence Startups in Africa - Nanalyze
"A recent paper published by Access Partnership and the University of Pretoria points out four important sectors where AI could greatly benefit Africa’s nations – agriculture, healthcare, public services, and financial services. The paper also highlights existing structural challenges that hamper AI development including little broadband coverage, inflexible education systems, and a lack of big data. Still, there are initiatives focused on overcoming these hurdles, and Google recently opened an AI research lab in Ghana, the first of its kind on the continent, led by ex-Facebook researcher Moustapha Cisse."

"Founded in 2014, Cape Town, South Africa startup Aerobotics has raised $4.8 million to develop a geospatial intelligence platform that enables pest and disease detection in tree crops using drones, satellite imagery, and of course artificial intelligence to interpret all the big data that’s being collected."
Nanalyze  AI  Africa  South-Africa  agriculture 
april 2019 by pierredv
AI Update: What Happens When a Computer Denies Your Insurance Coverage Claim? | Global Policy Watch Mar 2019
"Artificial intelligence is your new insurance claims agent. For years, insurance companies have used “InsurTech” AI to underwrite risk. But until recently, the use of AI in claims handling was only theoretical. No longer. The advent of AI claims handling creates new risks for policyholders, but it also creates new opportunities for resourceful policyholders to uncover bad faith and encourage insurers to live up to their side of the insurance contract."

"Now it is only a matter of time before insurers face pressure to use the available technology to deny claims as well.So what happens when a claim is denied?"

"If a policyholder prevails on a bad faith claim, it may be entitled to attorneys’ fees and punitive damages. Bad faith claims provide a counterweight to insurance companies’ information advantages, and can dramatically increase potential damages."

"The flip side of that complexity is that bad faith discovery may encourage early cooperation from the insurer. With their technology on the line, insurers may have a heightened incentive to pay what is due or otherwise settle before discovery for several reasons ..."
Covington  insurance  AI  law 
march 2019 by pierredv
DeepMind and Google: the battle to control artificial intelligence | 1843 April.May 2019
So far, Google has not interfered much with DeepMind. But one recent event has raised concerns over how long the company can sustain its independence.
Google  DeepMind  AI  TheEconomist 
march 2019 by pierredv
Tech Is Splitting the U.S. Work Force in Two - The New York Times, Feb 2019
"Despite all its shiny new high-tech businesses, the vast majority of new jobs are in workaday service industries, like health care, hospitality, retail and building services, where pay is mediocre."

"But automation is changing the nature of work, flushing workers without a college degree out of productive industries, like manufacturing and high-tech services, and into tasks with meager wages and no prospect for advancement. Automation is splitting the American labor force into two worlds. "

"Recent research has concluded that robots are reducing the demand for workers and weighing down wages, which have been rising more slowly than the productivity of workers. Some economists have concluded that the use of robots explains the decline in the share of national income going into workers’ paychecks over the last three decades."

"In a new study, David Autor of the Massachusetts Institute of Technology and Anna Salomons of Utrecht University found that over the last 40 years, jobs have fallen in every single industry that introduced technologies to enhance productivity. "
automation  technology  economics  NYTimes  employment  AI 
february 2019 by pierredv
AI Program Taught Itself How To 'Cheat' Its Human Creators | Zero Hedge
a columnist at TechCrunch highlighted a study that was presented at a prominent industry conference back in 2017. In the study, researchers explained how a Generative Adversarial Network - one of the two common varieties of machine learning agents - defied the intentions of its programmers and started spitting out synthetically engineered maps after being instructed to match aerial photographs with their corresponding street maps
AI  recognition  image-processing 
february 2019 by pierredv
Telecoms, media and digital economy prediction trends 2019 - AnalysysMason Jan 2019
Here are the headline predictions for 2019 from Analysys Mason Research:

5G: there will be many commercial 5G launches but consumers will not notice
IoT: the market for IoT connectivity will continue to develop and new models will appear
Enterprises: operators will buy IT operations to expand portfolios and there will be many SD-WAN launches
Small and medium-sized businesses: this market will streamline with DaaS and cloud apps
Homes: the battle for control of the emerging home ecosystem will intensify
Digital transformation: operators will offer mobile apps for customer care, and will push into edge computing and network slicing
Artificial intelligence: automation will lead operators to cut jobs and MSPs will capitalise on the needs of SMBs
Investor value: business simplification initiatives will enable telecoms operators worldwide to reduce headcount
AnalysysMason  prediction  IoT  5G  AI 
january 2019 by pierredv
Mind-reading devices can now access your thoughts and dreams using AI | New Scientist, Sep 2018
"We can now decode dreams and recreate images of faces people have seen, and everyone from Facebook to Elon Musk wants a piece of this mind reading reality"

"From an fMRI brain scan, Liu’s AI can say which of a selection of 15 different things a person was viewing when the scan was taken. For example, if someone was looking at a picture of a face, the AI can detect patterns in their scan that convince it to say “face”. Other options include birds, aeroplanes and people exercising, and the AI can call the correct category 50 per cent of the time."

Jack Gallant, UC Berkely: "When shown brain scans of someone watching a different YouTube video, the AI was able to generate a new movie of what it thought the person was viewing. The results are eerie outlines of the original, but still recognisable."

"Yukiyasu Kamitani at Japan’s Advanced Telecommunications Research Institute first showed in 2013 that it is possible to train an AI to detect the content of someone’s dreams, describing each in basic terms such as whether there was a male or female character, the objects included and details about the overall scene. Kamitani’s system has an accuracy of about 60 per cent."

"However, one big drawback of EEG is that there is so much unwanted noise to contend with. "

"The progress using AI with fMRI is causing people to rethink what EEG might be capable of."
NewScientist  AI  neuroscience  dreams  recognitioin  fMRI  EEG  ethics 
january 2019 by pierredv
The Global AI Race - Which Country Is Winning? - Nanalyze
Paraphrase of paper
Shoham et al., “The AI Index 2018 Annual Report”, AI Index Steering Committee, Human-Centered AI Initiative, Stanford University, Stanford, CA, December 2018.

Metrics: # papers and citations, patents, focus areas
Nanalyze  AI 
january 2019 by pierredv
Artificial intelligence is about to revolutionise warfare. Be afraid | New Scientist Sep 2018
"Sci-fi loves to depict military AIs as malign killer minds or robots. But the truth is more subtle and more terrifying – and it's happening right now"
NewScientist  AI  warfare  strategy 
december 2018 by pierredv
AI for Weather Forecasting – In Retail, Agriculture, Disaster Prediction, and More
"This article will look at how big data and machine learning are transforming weather forecasting and what it means for businesses and governments. In this article we’ll explore:

How companies and government agencies are using AI to improve weather forecasting (including IBM, Panasonic, and the US Government)
Sector-specific machine learning applications for improving business performance (including Retail, Agriculture, Transportation)

Weather forecasting is a strong fit for machine learning. The incredible volume of relevant information — historical data and real-time data — that can be analyzed is simply too great for any group of unaided humans to even begin to process on their own. "

"GE Current has installed smart street lights in several cities that can monitor things like light, humidity, and air quality."

"Panasonic has been working on its own weather forecasting model for years, and it stepped up its effort with the purchase of AirDat in 2013. The company makes TAMDAR, a speciality weather sensor installed on commercial airplanes. "

"According to IBM, 90 percent of crop losses are due to weather events and 25 percent of weather-related crop losses could be prevented by using predictive weather modeling."
TechEmergence  AI  ML  weather  forecasting  IBM  Panasonic  satellite  GE  IoT  NOAA  Monsanto  agriculturevideo 
november 2018 by pierredv
IBM’s machine argues, pretty convincingly, with humans - BBC News, Jun 2018
On a stage in San Francisco, IBM’s Project Debater spoke, listened and rebutted a human’s arguments in what was described as a groundbreaking display of artificial intelligence.

The machine drew from a library of “hundreds of millions” of documents - mostly newspaper articles and academic journals - to form its responses to a topic it was not prepared for beforehand.
AI  debate  discussion  conversation  BBC 
november 2018 by pierredv
XNOR | AI everywhere, on every device
via John Helm, Nov 2018
ex Allen Foundation people
AI 
november 2018 by pierredv
Artificial Intelligence | Global Catastrophic Risk Institute
"GCRI studies the human process of developing and governing AI, using risk analysis, social science, and the extensive knowledge we have gained from the study of other risks."

Provides list of papers
GCRI  risk-assessment  AI 
november 2018 by pierredv
Artificial Intelligence and Democracy | Philosophical Transactions of the Royal Society of London A: Mathematical, Physical and Engineering Sciences, Oct 2018
Via Gabor Molnar

Paul Nemitz

Abstract

Given the foreseeable pervasiveness of artificial intelligence (AI) in modern societies, it is legitimate and necessary to ask the question how this new technology must be shaped to support the maintenance and strengthening of constitutional democracy. This paper first describes the four core elements of today's digital power concentration, which need to be seen in cumulation and which, seen together, are both a threat to democracy and to functioning markets. It then recalls the experience with the lawless Internet and the relationship between technology and the law as it has developed in the Internet economy and the experience with GDPR before it moves on to the key question for AI in democracy, namely which of the challenges of AI can be safely and with good conscience left to ethics, and which challenges of AI need to be addressed by rules which are enforceable and encompass the legitimacy of democratic process, thus laws. The paper closes with a call for a new culture of incorporating the principles of democracy, rule of law and human rights by design in AI and a three-level technological impact assessment for new technologies like AI as a practical way forward for this purpose.
RoyalSociety  AI  democracy  politics 
october 2018 by pierredv
Self-driving car dilemmas reveal that moral choices are not universal - Nature, Oct 2018
"The largest ever survey of machine ethics1, published today in Nature, finds that many of the moral principles that guide a driver’s decisions vary by country. For example, in a scenario in which some combination of pedestrians and passengers will die in a collision, people from relatively prosperous countries with strong institutions were less likely to spare a pedestrian who stepped into traffic illegally."
NatureJournal  ethics  AI  morality  automobile 
october 2018 by pierredv
A Blueprint for the Future of AI - Brookings Oct 2018
Each of the papers in this series grapples with the impact of an emerging technology on an important policy issue, pointing out both the new challenges and potential policy solutions introduced by these technologies.
Brookings  AI  ML  robotics 
october 2018 by pierredv
Radio Frequency-Activity Modeling and Pattern Recognition (RF-AMPR) | SBIR.gov 2018
"OBJECTIVE: The PMW 120 Program Office desires a Radio Frequency Activity Modeling and Pattern Recognition (RF-AMPR) capability to perform pattern recognition, anomaly detection, and improved clustering of radio frequency (RF) signals. Specifically, it shall consist of a parametric RF classifier, a generative model of activity in the local electromagnetic environment, a machine learning-based anomaly detection method, and an RF data-clustering algorithm that classifies data that would otherwise be discarded by the parametric classifier."

"DESCRIPTION: Current automated RF data analysis and information discovery methods necessitate discarding significant volumes of sensor data as “non-analyzable”. This SBIR topic seeks to apply machine learning methodologies to better characterize this discarded data, enabling a more complete understanding of RF activity present in a specific environment."

"Anomaly classification shall include “known unknowns”, radio frequency events that are outliers of known classes, and “unknown unknowns”, anomalous RF events that represent new devices or activities."
SBIR  DoD  RF  spectrum  machine-learning  anomaly-classification  ML  AI 
october 2018 by pierredv
Deepfake Videos Are Getting Impossibly Good
"The new system was developed by Michael Zollhöfer, a visiting assistant professor at Stanford University, and his colleagues at Technical University of Munich, the University of Bath, Technicolor, and other institutions. Zollhöfer’s new approach uses input video to create photorealistic re-animations of portrait videos. These input videos are created by a source actor, the data from which is used to manipulate the portrait video of a target actor. So for example, anyone can serve as the source actor and have their facial expressions transferred to video of, say, Barack Obama or Vladimir Putin."
Gizmodo  video  AI 
october 2018 by pierredv
Colosseum: A Battleground for AI Let Loose on the RF Spectrum | 2018-09-15 | Microwave Journal
"In this article, we discuss the design and implementation of Colosseum, including the architectural choices and trades required to create an internet-based radio development and test environment of this scope and scale."
MicrowaveJournal  DARPA  Colosseum  AI  RF  simulation  mirror-worlds 
september 2018 by pierredv
Introduction to Local Interpretable Model-Agnostic Explanations (LIME) - O'Reilly Media
A technique to explain the predictions of any machine learning classifier.
By Marco Tulio RibeiroSameer SinghCarlos Guestrin
August 12, 2016

"In "Why Should I Trust You?" Explaining the Predictions of Any Classifier, a joint work by Marco Tulio Ribeiro, Sameer Singh, and Carlos Guestrin (to appear in ACM's Conference on Knowledge Discovery and Data Mining -- KDD2016), we explore precisely the question of trust and explanations."

"Because we want to be model-agnostic, what we can do to learn the behavior of the underlying model is to perturb the input and see how the predictions change."

"We used LIME to explain a myriad of classifiers (such as random forests, support vector machines (SVM), and neural networks) in the text and image domains."
ML  AI  neural-nets  explanation  prediction 
september 2018 by pierredv
Telecoms AI ecosystems: increasing automation in processes - Analysys Mason Jul 2018
Operators have been aiming to use automations driven by artificial intelligence (AI) to improve processes for a long time, but the effort involved has limited the number of instances in which intelligence can be applied. The introduction of AI ecosystems means that AI and other analytics techniques can be used to model processes and apply optimisation algorithms. Moreover, where processes require dynamic optimisation, machine learning (ML) and deep learning (DL) tools can automatically reoptimise processes as they change.
This report:

looks at the shift that is underway within the analytics market as telecom vendors develop telecoms-specific solutions
explains what the above shift is, the reasons behind the change and the implications for communications service providers (CSPs) and vendors
provides profiles of significant vendors in the market and a comparison of their approaches and tools.
AnalysysMason  AI  ML  telecoms 
september 2018 by pierredv
AI in telecommunications set to secure the ‘Augmented Human’ | Cambridge Consultants Feb 2018
"a new report, stating that the telecommunications sector will be an Artificial Intelligence (AI) pioneer, as many features of 5G networks and IoT networks will depend on AI techniques to reach their maximum potential"

"We expect to see data-led customisation: in healthcare with personalisation of treatments, in education with personalisation of teaching methods and curriculum, in entertainment with custom virtual reality games and maybe even in government. Of course, we also expect to see significant numbers of fully-autonomous vehicles on our roads and in our corridors transporting people and goods to where they are required."
ambridgeConsultants  AI  telecoms  spectrum  5G  IoT 
september 2018 by pierredv
Google Makes Its Special A.I. Chips Available to Others - The New York Times
"A few years ago, Google created a new kind of computer chip to help power its giant artificial intelligence systems. These chips were designed to handle the complex processes that some believe will be a key to the future of the computer industry.

On Monday, the internet giant said it would allow other companies to buy access to those chips through its cloud-computing service. Google hopes to build a new business around the chips, called tensor processing units, or T.P.U.s."
NYTimes  Google  AI  chips 
september 2018 by pierredv
The Path of Robotics Law by Jack M. Balkin :: SSRN May 2015
Via Blake Reid

"Lawrence Lessig's famous dictum that "Code is Law" argued that combinations of computer hardware and software, like other modalities of regulation, could constrain and direct human behavior. Robotics and AI present the converse problem. Instead of code as a law that regulates humans, robotics and AI feature emergent behavior that escapes human planning and expectations. Code is lawless. "
robotics  AI  law  cyberlaw 
september 2018 by pierredv
Army turns to artificial intelligence to counter electronic attacks - SpaceNews.com aug 2018
"A team of eight engineers from Aerospace Corp. won a $100,000 Army prize by correctly detecting and classifying the greatest number of radio frequency signals using a combination of signal processing and artificial intelligence algorithms"
SpaceNews  AI  ML  spectrum  USArmy  competition  signal-classification 
august 2018 by pierredv
RSPG to examine role of machine learning | PolicyTracker aug 2018
"There is growing interest in the application of artificial intelligence (AI) to spectrum: Google thinks it could replace propagation modelling; a leading consultancy has described the telecoms sector as a "perfect opportunity" and AI will be one of the forthcoming study areas for the EU's spectrum advisory group. "
PolicyTracker  AI  ML  RSPG  spectrum 
august 2018 by pierredv
The rise of 'pseudo-AI': how tech firms quietly use humans to do bots' work | Technology | The Guardian
"It’s hard to build a service powered by artificial intelligence. So hard, in fact, that some startups have worked out it’s cheaper and easier to get humans to behave like robots than it is to get machines to behave like humans.""In 2016, Bloomberg highlighted the plight of the humans spending 12 hours a day pretending to be chatbots for calendar scheduling services such as X.ai and Clara. The job was so mind-numbing that human employees said they were looking forward to being replaced by bots. "
TheGuardian  AI  employment  start-ups  technology  business 
july 2018 by pierredv
Artificial Intelligence and its Implications for Income Distribution and Unemployment – Economic and Policy Implications of AI, Blog Post #3 | The Technology Policy Institute
"The following is a summary of Artificial Intelligence and its Implications for Income Distribution and Unemployment by Antonin Korinek and Joseph E. Stiglitz. This paper was presented at the Technology Policy Institute Conference on The Economics and Policy Implications of Artificial Intelligence, February 22, 2018."

"In Artificial Intelligence and its Implications for Income Distribution and Unemployment, Anton Korinek and Joseph E. Stiglitz focus on a key economic challenge associated with increased use of AI: its effects on income distribution in the context of the future of work. They discuss four cases in which AI innovations are more likely to be substitutes rather than complements for human labor under varying market conditions, and the potential for Pareto improvement[1] in the economic outcomes of both workers and innovators."

"They note only two market and societal contexts in which AI innovation might result in Pareto improvement, though only one is realistically possible. Though they call for policies that would support a move toward Pareto improvement – specifically to ensure redistributing income or subsidizing workers’ wages – the authors close by acknowledging the long and complex road ahead. "
TPI  AI  employment  economics 
june 2018 by pierredv
AI researchers allege that machine learning is alchemy | Science | AAAS, May 2018
Via John Helm

"Speaking at an AI conference, Rahimi charged that machine learning algorithms, in which computers learn through trial and error, have become a form of "alchemy." Researchers, he said, do not know why some algorithms work and others don't, nor do they have rigorous criteria for choosing one AI architecture over another. Now, in a paper presented on 30 April at the International Conference on Learning Representations in Vancouver, Canada, Rahimi and his collaborators document examples of what they see as the alchemy problem and offer prescriptions for bolstering AI's rigor."

"The issue is distinct from AI's reproducibility problem... It also differs from the "black box" or "interpretability" problem in machine learning"

"Without deep understanding of the basic tools needed to build and train new algorithms, he says, researchers creating AIs resort to hearsay, like medieval alchemists."
ML  machine-learning  AI  ScienceMag 
may 2018 by pierredv
House of Lords Select Committee publishes report on the future of AI in the UK | Covington Global Policy Watch Apr 2018
"Reflecting evidence from 280 witnesses from the government, academia and industry, and nine months of investigation, the UK House of Lords Select Committee on Artificial Intelligence published its report “AI in the UK: ready, willing and able?” on April 16, 2018 (the Report). The Report considers the future of AI in the UK, from perceived opportunities to risks and challenges. In addition to scoping the legal and regulatory landscape, the Report considers the role of AI in a social and economic context, and proposes a set of ethical guidelines. This blog post sets out those ethical guidelines and summarises some of the key features of the Report."

"Access to Data
One of the key concerns set out in the Report is that all companies should have “fair and reasonable access to data”. The Committee notes the possibility that the ‘Big Tech’ companies may use network effects to build up large proprietary datasets which are difficult to match. In response, the Committee suggests ethical, data protection and competition frameworks, including a review of the use of data by the Competition and Markets Authority."
Covington  AI  UK 
may 2018 by pierredv
May 2018 - AI Heads to Space with a Constellation of Use Cases | Via Satellite Apr 2018
"Though the technology is just starting to roll out, Artificial Intelligence is applicable to many aspects of the satellite ecosystem, including system manufacturing, in-orbit management and image processing. From optimizing satellite builds on the factory floor to accurately determining the GDP of an entire nation, automated, scalable algorithms offer the space ecosystem rich potential for both an operational sea-change and a whole new set of applications."

"It’s important to note that AI may represent enormous possibilities for process improvement, but its deployment is not without challenges. One of the main issues on the manufacturing front is the intense amount of customization that’s required. No two satellites are identical, despite attempts to standardize products and systems."

In orbit: "... SES, whose fleet of 70 satellites each generates continuous health and status data. The operator is working with IBM Watson to use a real-time streaming AI to look at that data lake and increase reaction time to correct and normalize satellite operations when fluctuations in the telemetry occur."

"AI however opens up a new chapter in EO by lending scale, automated image tagging and sorting, and map correlations — which, taken together, can be used for entirely new applications, such as understanding the economic state of a third-world nation."
ViaSatellite  AI  space  satellite 
april 2018 by pierredv
Defense intelligence chief: ‘A lot of technology remains untapped’ - SpaceNews.com Apr 2018
"The most promising AI effort the Pentagon has going now is Project Maven. Military analysts are using Google-developed AI algorithms to mine live video feeds from drones. With machine learning techniques, software is taught to find particular objects or individuals at speeds that would be impossible for any human analyst."

"... DoD ramps up AI efforts. Defense procurement chief Ellen Lord said the Pentagon will start bringing together AI projects that already exist but do not necessarily share information or resources. "
SpaceNews  AI  Google  DoD  remote-sensing  intelligence-gathering 
april 2018 by pierredv
Do Algorithms Rule the World? Algorithmic Decision-Making and Data Protection in the Framework of the GDPR and Beyond – Economic and Policy Implications of AI, Blog Post #2 | The Technology Policy Institute Apr 2018
"a summary of Do Algorithms Rule the World? Algorithmic Decision-Making and Data Protection in the Framework of the GDPR and Beyond by Dr. Maja Brkan. This paper was presented at the Technology Policy Institute Conference on The Economics and Policy Implications of Artificial Intelligence, February 22, 2018."

"Article 22 of the GDPR gives individuals the right not to be subject to a decision based solely on automated processing and prohibits automated decision-making that produces binding effects."
TPI  privacy  AI  GDPR 
april 2018 by pierredv
▶ Deep Reinforcement Learning - VideoLectures.NET
Description
In this tutorial I will discuss how reinforcement learning (RL) can be combined with deep learning (DL). There are several ways to combine DL and RL together, including value-based, policy-based, and model-based approaches with planning. Several of these approaches have well-known divergence issues, and I will present simple methods for addressing these instabilities. The talk will include a case study of recent successes in the Atari 2600 domain, where a single agent can learn to play many different games directly from raw pixel input.
VideoLectures  video  deep-learning  lectures  AI 
april 2018 by pierredv
Twitter Bots: An Analysis of the Links Automated Accounts Share | Pew Research Center - Apr 2018
"An estimated two-thirds of tweeted links to popular websites are posted by automated accounts – not human beings"

"This analysis finds that the 500 most-active suspected bot accounts are responsible for 22% of the tweeted links to popular news and current events sites over the period in which this study was conducted. By comparison, the 500 most-active human users are responsible for a much smaller share (an estimated 6%) of tweeted links to these outlets."
Pew  Twitter  AI  bots  socialmedia  socialnetworking 
april 2018 by pierredv
Deep learning - Yann LeCun, Yoshua Bengio & Geoffrey Hinton | Nature May 2015
Abstract
Deep learning allows computational models that are composed of multiple processing layers to learn representations of data with multiple levels of abstraction. These methods have dramatically improved the state-of-the-art in speech recognition, visual object recognition, object detection and many other domains such as drug discovery and genomics. Deep learning discovers intricate structure in large data sets by using the backpropagation algorithm to indicate how a machine should change its internal parameters that are used to compute the representation in each layer from the representation in the previous layer. Deep convolutional nets have brought about breakthroughs in processing images, video, speech and audio, whereas recurrent nets have shone light on sequential data such as text and speech.
machine-learning  ML  AI  NatureJournal 
april 2018 by pierredv
Machine Learning for Performance Prediction in Mobile Cellular Networks - IEEE Computational Intelligence Magazine ( Volume: 13, Issue: 1, Feb. 2018 )
Janne Riihijarvi ; Petri Mahonen

Abstract:
In this paper, we discuss the application of machine learning techniques for performance prediction problems in wireless networks. These problems often involve using existing measurement data to predict network performance where direct measurements are not available. We explore the performance of existing machine learning algorithms for these problems and propose a simple taxonomy of main problem categories. As an example, we use an extensive real-world drive test data set to show that classical machine learning methods such as Gaussian process regression, exponential smoothing of time series, and random forests can yield excellent prediction results. Applying these methods to the management of wireless mobile networks has the potential to significantly reduce operational costs while simultaneously improving user experience. We also discuss key challenges for future work, especially with the focus on practical deployment of machine learning techniques for performance prediction in mobile wireless networks.
machine-learning  ML  automation  AI  IEEE  cellular  spectrum 
april 2018 by pierredv
Analyzing Financial Forecasting Models Using AI - Nanalyze - April 2018
A startup called Visible Alpha wants to change the way that analysts construct financial forecasting models by making the data they need (and plenty of data they didn’t know they needed) all readily available at the click of a button.
Nanalyze  investing  NLP  language-recognition  automation  AI 
april 2018 by pierredv
Digital Reasoning - AI That Understands You - Nanalyze - Apr 2018
We first came across Digital Reasoning in our article on Artificial Intelligence (AI) as a Service and “Core AI” in which we noted that their technology, Synthesys, is exceptionally good at understanding what people are actually saying when they use analogies.
Nanalyze  AI  automation  language-recognition  compliance 
april 2018 by pierredv
CRFS publishes White Paper on Machine Learning — CRFS, Dec 2017
Via Dale Hatfield, March 2018

"Machine Learning and RF Spectrum Intelligence Gathering"

"Many applications that are central to RF spectrum intelligence gathering require some sort of pattern recognition. For example, to classify a signal by type we need to identify the particular pattern associated with the modulation, while to recognise that there is an interesting signal present in received data, we need to distinguish between pattern and noise.

In this White Paper, we explore how machine learning techniques can be applied to these applications of signal classification and anomaly detection to deliver faster and more effective performance to customers."
CRFS  signal-processing  enforcement  AI  ML  machine-learning  spectrum 
april 2018 by pierredv
Need to make a molecule? Ask this AI for instructions - Nature Mar 2018
"Researchers have developed a ‘deep learning’ computer program that produces blueprints for the sequences of reactions needed to create small organic molecules, such as drug compounds. The pathways that the tool suggests look just as good on paper as those devised by human chemists."

"Segler and his team tested the pathways that the program threw up in a double-blind trial, to see whether experienced chemists could tell the AI’s synthesis pathways from those devised by humans. They showed 45 organic chemists from two institutes in China and Germany potential synthesis routes for nine molecules: one pathway suggested by the system, and another devised by humans. The chemists had no preference for which was best."

"Segler’s tool is different because it learns from the data alone and does not need humans to input rules for it to use."
AI  chemistry  NatureJournal 
march 2018 by pierredv
U.S. GAO - Technology Assessment: Artificial Intelligence: Emerging Opportunities, Challenges, and Implications - March 2018
The Comptroller General convened the Forum on AI to consider the policy and research implications of AI’s use in 4 areas with the potential to significantly affect daily life—cybersecurity, automated vehicles, criminal justice, and financial services. The forum highlighted the fact that AI will have far-reaching effects on society—even if AI capabilities stopped advancing today. We looked at the prospects for AI in the near future, and identified areas where changes in policy and research may be needed.
GAO  AI 
march 2018 by pierredv
Artificial intelligence faces reproducibility crisis | Science
"Summary
The booming field of artificial intelligence (AI) is grappling with a replication crisis, much like the ones that have afflicted psychology, medicine, and other fields over the past decade. Just because algorithms are based on code doesn't mean experiments are easily replicated. Far from it. Unpublished codes and a sensitivity to training conditions have made it difficult for AI researchers to reproduce many key results. That is leading to a new conscientiousness about research methods and publication protocols. Last week, at a meeting of the Association for the Advancement of Artificial Intelligence in New Orleans, Louisiana, reproducibility was on the agenda, with some teams diagnosing the problem—and one laying out tools to mitigate it."

"Odd Erik Gundersen, a computer scientist at the Norwegian University
of Science and Technology in Trondheim, reported the results of a survey of 400 algorithms presented in papers at two top AI conferences in the past few years. He found that only 6% of the presenters shared the algorithm’s code. Only a third shared the data they tested their algorithms on, and just half shared “pseudocode”—a limited summary of an algorithm."
ScienceMag  reproducibility  replication  AI 
february 2018 by pierredv
He Predicted The 2016 Fake News Crisis. Now He's Worried About An Information Apocalypse.
That future, according to Ovadya, will arrive with a slew of slick, easy-to-use, and eventually seamless technological tools for manipulating perception and falsifying reality, for which terms have already been coined — “reality apathy,” “automated laser phishing,” and "human puppets."
politics  AI  trends  news  journalism  socialmedia 
february 2018 by pierredv
Can Washington Be Automated? - POLITICO Magazine - Feb 2018
Via Blake Reid

"This kind of data-crunching might sound hopelessly wonky, a kind of baseball-stats-geek approach to Washington. But if you’ve spent years attempting to make sense of the Washington information ecosystem—which can often feel like a swirling mass of partially baked ideas, misunderstandings and half-truths—the effect is mesmerizing. FiscalNote takes a morass of documents and history and conventional wisdom and distills it into a precise serving of understanding, the kind on which decisions are made."

"We aren’t far from a future where public commenting on regulations—the process for individual American citizens to offer feedback to their elected government—comes down to a bot vs. bot fight."
automation  politics  Politico  lobbying  AI 
february 2018 by pierredv
The Radio Frequency Spectrum + Machine Learning = A New Wave in Radio Technology
"The radio frequency spectrum is becoming increasingly crowded and a new DARPA program will examine how leading-edge machine learning can help understand all the signals in the crowd"

“What I am imagining is the ability of an RF Machine Learning system to see and understand the composition of the radio frequency spectrum – the kinds of signals occupying it, differentiating those that are ‘important’ from the background, and identifying those that don’t follow the rules,” said Tilghman.

The RFMLS program features four technical components that would integrate into future RFML systems:

Feature Learning: ...
Attention and Saliency: ...
Autonomous RF Sensor Configuration: ...
Waveform Synthesis: ...
DARPA  AI  RF  spectrum  machine-learning 
november 2017 by pierredv
Why should I trust you? Explaining the predictions of any classifier | the morning paper
Summary of “Why Should I Trust You? Explaining the Predictions of Any Classifier Ribeiro et al., KDD 2016
machine-learning  AI  automation 
november 2017 by pierredv
EU citizens might get a 'right to explanation' about the decisions algorithms make, Splinter Jul 2016
"Late last week, though, academic researchers laid out some potentially exciting news when it comes to algorithmic transparency: citizens of EU member states might soon have a way to demand explanations of the decisions algorithms about them. In April, the EU approved a data protection law called the General Data Protection Regulation (GDPR)."

"In a new paper, sexily titled "EU regulations on algorithmic decision-making and a 'right to explanation,'" Bryce Goodman of the Oxford Internet Institute and Seth Flaxman at Oxford's Department of Statistics explain how a couple of subsections of the new law, which govern computer programs making decisions on their own, could create this new right."

"While the new provision may seem great at first glance, the word "solely" makes the situation a little more slippery, says Ryan Calo ... Calo explained over email how companies that use algorithms could pretty easily sidestep the new regulation. "All a firm needs to do is introduce a human—any human, however poorly trained or informed—somewhere in the system," Calo said."
EU  GDPR  algorithms  automation  AI 
november 2017 by pierredv
Computers that Code Themselves Using AI - Nanalyze
"Before we delve into this notion of computers that can code themselves, we need to consider that traditional coding is becoming less and less useful. Take as an example the latest flavor of Google’s AlphaGo algorithm. It was told what the rules of a game of Go looked like, and then it was told not to lose. The algorithm then created its own method of playing Go as opposed to being “trained” by watching other people play"

"The way that your brain works isn’t something that you could create using today’s best coding languages, and neither will AI. With that said, there are some startups out there that are developing technology that can write and rewrite itself. "

"Founded in 2013, Massachusetts startup Gamalon has taken in around $12 million in funding to develop a system that learns “orders of magnitude faster” and with “orders of magnitude less training data” compared to the more traditional machine learning algorithms of today."

"The fact that the computer does something so exponentially well that it renders a team of coders useless is the same net effect as a computer that can code itself. The end result is that we don’t need so many coders."
nanalyze  automation  software  AI  development-sw  employment 
november 2017 by pierredv
Reshaping Business With Artificial Intelligence - MIT Sloan, Sep 2017
Via TPI

"Disruption from artificial intelligence (AI) is here, but many company leaders aren’t sure what to expect from AI or how it fits into their business model. Yet with change coming at breakneck speed, the time to identify your company’s AI strategy is now. MIT Sloan Management Review has partnered with The Boston Consulting Group to provide baseline information on the strategies used by companies leading in AI, the prospects for its growth, and the steps executives need to take to develop a strategy for their business."
data  business  AI  automation  MIT  strategy 
november 2017 by pierredv
Machine learning still needs data scientists to optimise results - Comments - Content | EPiServer Site
"Machine learning is not a 'one-size-fits-all' technology, but a growing library of technologies that need to be understood and deployed correctly to achieve meaningful results."

Good survey of techniques and algorithms
AnalysysMason  MachineLearning  AI  automation  Communications  algorithms 
october 2017 by pierredv
One year later, Microsoft AI and Research grows to 8k people in massive bet on artificial intelligence – GeekWire, Sep 2017
"One way to measure Microsoft’s AI bet: In its first year of operation, the AI and Research group has grown by 60 percent — from 5,000 people originally to nearly 8,000 people today — through hiring and acquisitions, and by bringing aboard additional teams from other parts of the company."
AI  Microsoft  GeekWire 
september 2017 by pierredv
8 Satellite Data Startups Doing Geospatial Analysis - Nanalyze - Nanalyze
Capella Space - CEO "emphasizes that while his company builds satellites, it considers itself an information business"

Ursa Space Systems - "main products include a weekly time series on oil stocks down to the tank level"

SpaceKnow - "business intelligence based on imagery [plus] offers a range of other products, based on its machine-learning platform"

TellusLabs - " combining satellite imagery with other data such as local weather and crop conditions, its machine-learning algorithms ..."

UrtheCast - "does a little bit of everything in satellite data ... makes most of its money from selling satellite imagery and even video from its own satellites and space-based cameras"
satellite  geospatial-analysis  NewSpace  AI  MachineLearning  remote-sensing  EO 
september 2017 by pierredv
Smart buildings predict when critical systems are about to fail | New Scientist issue 3110, Jan 2017
"They trained a machine learning algorithm on data from the first half of 2015, looking for differences in the readings of similar appliances. They then tested it on data from the second half of the year – could it predict faults before they happened? The system predicted 76 out of 124 real faults, including 41 out of 44 where an appliance's temperature rose above tolerable levels, with a false positive rate of 5 per cent"

See https://arxiv.org/ftp/arxiv/papers/1701/1701.03633.pdf

"Finnish start-up Leanheat puts a wireless temperature, humidity and pressure sensor into apartments to remotely control heating and monitor appliance health. Its system is now installed in nearly 400 apartment blocks, says chief executive Jukka Aho."Once we had these sensors in place, very quickly there was evidence that buildings were not controlled optimally," he says. Instead of adjusting heating simply based on the outside temperature, Leanheat's models take into account how the weather has changed. Has the temperature fallen to zero from 10 degrees – or risen from 10 below?"
"US-based start-up Augury is installing acoustic sensors in machines to listen for audible changes in function and spot potentially imminent failures. CEO Saar Yoskovitz says Augury has 'diagnosed' machines in hospitals, power plants, data centres and a university campus."
NewScientist  building  architecture  RF-MirrorWorlds  prediction  AI  machine-learning 
may 2017 by pierredv
The world in 2076: Machines outsmart us but we're still on top | New Scientist issue 3100, 19 Nov 2016
Article has good arguments against the singularity by Toby Walsh

1. The "fast-thinking dog" argument, quoting Steven Pinker, "Sheer processing power is not a pixie dust that magically solves all your problems."

2. Anthropocentric argument. "There is no reason to suppose that human intelligence is a tipping point, that once passed allows for rapid increases in intelligence."

3. The "diminishing returns" argument. "The idea of a technological singularity supposes that improvements to intelligence will be by a relative constant multiplier, each generation getting some fraction better than the last. However, the performance of most of our AI systems has so far been that of diminishing returns. There are often lots of low-hanging fruit at the start, but we then run into difficulties when looking for improvements."

4. The "limits of intelligence" argument. "... AI may well run into some fundamental limits. Some of these may be due to the inherent uncertainty of nature. No matter how hard we think about a problem, there may be limits to the quality of our decision-making."

5. The "computational complexity" argument. "There are many computational problems for which even exponential improvements are not enough to help us solve them practically. "
NewScientist  AI  singularity 
may 2017 by pierredv
« earlier      
per page:    204080120160

Copy this bookmark:





to read