Artificial Intelligence, Automation, and Robotics

Redshift: 3 Ways Robots Can Shape the Future of Architecture—On Earth and Beyond

2018-04-30T10:15:00.000-07:00

This Redshift article - 3 Ways Robots Can Shape the Future of Architecture - On Earth and Beyond - examines the future of robots with benefit of insights gleaned from Towards a Robotic Architecture, a text coedited by Dr. Mahesh Daas (KU's Dean of Architecture) and Andrew John Witt (assistant professor of Digital Practice within Temple University’s Division of Architecture and Environmental Design). Enjoy!

Science: Europe to Devote €1.5 Billion to Artificial Intelligence Research

2018-04-30T07:13:00.000-07:00

To distinguish their research efforts, European AI research will be guided by a focus on ethics in contrast to the "permissionless innovation" approach taken by the U.S. and China.

For the full article, see: http://www.sciencemag.org/news/2018/04/15-billion-artificial-intelligence-research-europe-pins-hopes-ethics

NSF Cyber-Physical Systems Grant Solicitation Released

2018-02-16T08:25:00.000-08:00

NSF has released its 2018 CPS solicitation. The program funds projects for engineered systems that are built from and depend upon, the seamless integration of computation and physical components. The CPS program aims to develop the core research needed to engineer complex cyber-physical systems, some of which may also require dependable, high-confidence, or provable behaviors. Core research areas of the program include control, data analytics, autonomy, design, information management, internet of things (IoT), mixed initiatives including human-in- or on-the-loop, networking, privacy, real-time systems, safety, security, and verification.

Small projects: $500k/3 yrs
Medium projects: $500k-1M/3 yrs
Frontier projects $1M-7M/4-5 yrs

Proposal Deadline: May 8, 2018

Links:
Program Description
Solicitation

God 2.0?

2017-11-20T13:43:00.000-08:00

In this Wired article, former Google/Uber wunderkind Anthony Levandowski envisions the future of AI as a "for-prophet" not-for-profit venture. Levandowski seeks converts who embrace the tenets of his newly founded religion Way of the Future. It's goal - to herald and shape mankind's new relationship with our inevitable AI divinity. When it comes to AI, it seems no job is safe.

The Business of Artificial Intelligence

2017-11-14T12:30:00.001-08:00

We've spent a fair number of pixels prognosticating the future of AI. This article from the Harvard Business Review refreshingly discusses what AI can do for you lately. The good news is if you use AI to guide your food choices, then since 2016, your risk of accidentally consuming a chihuahua's head has decreased dramatically. Enjoy!

EVENT: A New Mode of Orientation: Planetary Cognitive Ecologies.

2017-10-20T14:13:00.001-07:00

Presented By: N. Katherine Hayles, Distinguished Professor Emerita, Literature, Duke University
What: Special Speaker for the Hall Center Seminar on "The Posthuman?"
Where: KU Alumni Center
When: Monday, November 13, 2017, 7:30-9 p.m.
Suggested Background Reading: How We Became Posthuman (1999)

Questions? 864-2519 or relliott@ku.edu

Hosted by: Hall Center for the Humanities.

EVENT: Coffee @ The Commons

2017-10-03T11:40:00.001-07:00

AI Raises the Bar

2017-07-13T09:19:00.000-07:00

Stanford University Junior Joshua Browder's AI Legal chatbot DoNotPay has purportedly helped defeat 375,000 parking tickets over the last two years and is now available nationwide, according to The Verge. TechCrunch reports that IBM's Watson technology has been offered to DoNotPay for free to enable users to express questions using natural language.

Should attorneys worry that AI will replace them? Not according to the ABA's Litigation News. Indeed, although large firms are retaining various AI systems for tasks such as due diligence, lease abstraction, and contract analysis, at least one firm characterizes those systems as supplemental tools to help attorneys produce results faster, learn more quickly, and improve - not to replace them.

Which raises the question - is the potential to generate more legal work truly limitless?

At least from an output perspective, DoNotPay would seem to suggest it might be. Below is a hastily-generated sample. Interestingly, the program doesn't yet appear to provide instructions for how to proceed when all those camera maintenance records arrive.

To Whom it May Concern,

I received a letter claiming I committed a moving traffic violation in lawrence kansas on 7/13/2017.

As per the instructions, I am writing to plead 'not guilty' to this charge. Although this option is said to result in this matter going to court; it is my suggestion that the charges simply be dropped. This suggestion comes out of respect for tax payers, and my request that their hard earned money not be wasted in such proceedings.

As there is no evidence of my involvement with this alleged 'crime', as well as the fact that I am not granted my 6th amendment right to face my 'accuser' (a camera); I see no way the government could prove my guilt beyond a reasonable doubt. I also see find no legal requirement for me to implicate someone else in this process, as it is the government's responsibility to prove a person's guilt. It is also my 5th amendment right to remain silent on the matter.

If it is the government's decision to move forward in this matter, I would request copies of any evidence the prosecution may have of my involvement in the 'offense'; as well as, all maintenance records for the camera(s) involved.

Sincerely,

Test Case

Surveyed Truckers Skeptical Self-Driving Trucks Will Happen Soon.

2017-06-26T08:09:00.000-07:00

From Mark Zuckerberg on Facebook, during a visit to Iowa 80, The World's Largest Truck Stop:

"I asked a number of truckers what they think about self-driving cars and trucks and what they think about the future. Everyone I met was skeptical self-driving trucks would replace jobs for different reasons. Some thought it would be impossible to pack all the sensors you need to deal with things like weather into trucks. Others thought computers could handle the interstate but not the last mile to the store. And some truckers think we'll end up with something like autopilot on planes -- with trucks driving themselves with people in the cab. From all the research I've seen, I'm confident we'll solve these problems. But it's interesting that people in the industry don't believe this will happen soon."

Link to full post and comments here.

EVENT: Frontiers of Large-Scale Data Analytics

2017-06-09T14:42:00.000-07:00

What: Large-Scale Data Analytics Workshop:

Where: Nichols Hall, Room 317

When: June 16th, 2017 at 9:00 a.m.

RSVP requested: Sohaib Kiani (sohaib.kiani at ku.edu)

Event Flyer: Below

The Speed of Thought

2017-05-16T14:10:00.000-07:00

In the NY Times Magazine article The Great A.I. Awakening, Gideon Lewis-Kraus relates the tale of the development of the AI implementation of Google Translate. The previous version of Translate, which had been running for 10 years, was based on a symbolic A.I. model. That platform required exhaustive human-managed data sets and rules. By contrast, the new version was built on a machine-learning neural-network platform. Time taken to complete the new system--nine months.

Here's a Japanese translation of Hemingway's The Snows of Kilimanjaro translated back to English as compared with the original:

NO. 1:
Kilimanjaro is a snow-covered mountain 19,710 feet high, and is said to be the highest mountain in Africa. Its western summit is called the Masai “Ngaje Ngai,” the House of God. Close to the western summit there is the dried and frozen carcass of a leopard. No one has explained what the leopard was seeking at that altitude.

NO. 2:
Kilimanjaro is a mountain of 19,710 feet covered with snow and is said to be the highest mountain in Africa. The summit of the west is called “Ngaje Ngai” in Masai, the house of God. Near the top of the west there is a dry and frozen dead body of leopard. No one has ever explained what leopard wanted at that altitude.

Can you tell which is the Google translation?

Here's what the translation would have looked like using Translate the day before the switch to the new version:

Kilimanjaro is 19,710 feet of the mountain covered with snow, and it is said that the highest mountain in Africa. Top of the west, “Ngaje Ngai” in the Maasai language, has been referred to as the house of God. The top close to the west, there is a dry, frozen carcass of a leopard. Whether the leopard had what the demand at that altitude, there is no that nobody explained.

Change is coming - far faster than anyone could have predicted.

Setting Standards

2017-05-01T11:22:00.001-07:00

"Is that SAE 30 weight?"

I spent a lot of time working on industry standards for heavy equipment. At times it seemed kind of silly--"What should we call this thing that attaches to a loader?" or "What is the definition of bucket capacity?" But in the big scheme of things standards make the world go round. If you didn't have standard times, or standard bolt sizes, or a common definition of a meter, all industry would be stymied. International standards for equipment safety (like rollover protection or seat belts) define a consensus view of minimum acceptable design practices. If you take it a step further--standards define some level of due diligence from a product liability perspective. Building a product that doesn't meet standard is generally considered negligent.

Well, the AI industry is starting to work on AI standards and best practices. The "Partnership on Artificial Intelligence to Benefit People and Society" is a group that includes Amazon, Google, Microsoft, Apple, the ACLU, and Facebook. Their kickoff Board of Trustees meeting is this month.
www.partnershiponai.org

I would strongly encourage you to track what the Partnership does and get involved. Standards groups are led by industry and reflect their market and product concerns. But standards groups need the expertise of non-industry scientists who deeply understand technology, social science, and system behavior. The issues that must be addressed are huge and wide-ranging. Participating in standards development can help your research program focus on industry-relevant needs, you can make great connections to potential partners and collaborators. You might even get to draft a new ISO standard warning label:

You know it's serious when Congress holds a hearing ...

2017-04-26T12:50:00.002-07:00

When technology develops to the point where real-world applications start to have an impact, it usually ends up in front of Congress for a hearing. This shows everyone that Congress is paying attention to cutting-edge issues and it plays an important role in helping to define Federal interests and necessary action. The first ever Congressional hearing on artificial intelligence was held November 30, 2016 by the Commerce, Science, and Transportation subcommittee.
AI Hearing

The Federal government has some key objectives relative to AI:
1) Make sure that the U.S. is able to take competitive advantage of the technology. We hate to get behind in anything since 1957.
2) Related to #1--remove any unintentional barriers that hinder the new technology from being implemented.
3) Opposite to #2--make sure we properly limit/regulate/address negative consequences of the new technology through the establishment of new policy

Alice Bean should weigh in on how the Federal government is likely to develop policy around AI technology. She served a State Department fellowship focused on science and policy.

Right now Congress is in a facilitative position. They want to enable adoption of self-driving cars and the expanded use of UAV's. They are being careful about having the FAA and NHTSA look at safety issues, however they are still in the optimistic zone touting reduced highway deaths, expanded economy, new high-tech jobs, less pollution, more apple pie. Sen. Nelson finally raised the question of job displacement about minute 32 of the hearing. Dr. Horvitz notes that we are "at an inflection point" right now with dramatically expanding AI capability. Benefits and impacts will accelerate. Dr. Moore suggests that we need a Scholarship for Service program for AI like we have for cybersecurity. Mr. Brockman notes that AI moves from basic research to products in a matter of months, not years. The panel's Q&A starts at about 1 hour.

This kind of discussion is often the lead-in to new funding opportunities--stay tuned to see how agencies respond.

Artificial Intelligence: Here to Help

2017-04-19T07:27:00.001-07:00

A lot of recent media reports have been highlighting negative impacts of artificial intelligence systems on society (loss of jobs, super AI, killer robots). Those things certainly get your attention. However a new report (March 2017) by the Computing Community Consortium describes the potential benefits of AI. "Artificial Intelligence for Social Good" is a summary of a workshop that was held last year with examples of the promise of AI in urban computing, sustainability, health care, and public welfare. The report describes current case study examples and offers tantalizing glimpses of a rosy future--"The easier it becomes for people to move about, the more vibrant our urban areas will be; likewise, the more fruitful the social and economic interactions that take place inside them will be." Really?

Sorry, I am a skeptic. How do I get from self-driving Uber to more equitable access to transportation in urban areas? How does easier access to transportation lead to "a more vibrant" urban area? I think this is really the key point of where we are right now in research needs--the great gap in knowledge is to understand how AI/robotic tech will actually deploy into and impact society. Could you simulate right now what the impact of 10% penetration of self-driving, publically accessible vehicles into the Lawrence Metropolitan Statistical Area would be? Who would ride in them? What parts of the local economy would be affected? Who would own them? How would it affect legacy transit modes--buses, bikes, personal cars, taxis? What demands would it put on public services? The report says we need more pilot studies. Any takers?

EVENT: Artificial Intelligence - A Cross-Disciplinary Forum

2017-04-17T12:16:00.002-07:00

What: Six short, interactive presentations on AI impacts from KU speakers in a variety of fields.
When: 10 a.m. to Noon, Friday August 12, 2017
Where: The Commons on KU's Lawrence Campus

Stay tuned - Program details to be announced.

Robot-proof Jobs

2017-04-14T07:56:00.000-07:00

Interesting series this week on NPR's Marketplace with David Brancaccio about jobs that may be "robot-proof."
NPR Marketplace Robot-Proof Jobs

He traveled around the country interviewing people that are doing tasks that will be hard to automate (like music composer for example). His background information was drawn from the McKinsey report on potential for automation. This is a work in progress and the page has links to Tableau datasets and a cool "suitable for framing" poster of jobs and automation.
McKinsey Global Report

Bottomline: McKinsey suggests about 45% of paid job functions could be replaced by AI/robotics technology that is available today. They note that about 60% of occupations will likely see some amount of AI/robotic impact.

So at the end of this morning's NPR segment, Brancaccio mentioned an interesting concept--since people won't be able to be paid for their work we will all need to own a robot or two that will work for us and thus we will earn a "living" from the productive return on our capital investments. A robot would be your surrogate worker.

EVENT: Design Automation Meets Artificial Intelligence

2017-04-12T13:50:00.000-07:00

Who: Dr. Michael Hsio, Professor, IEEE Fellow, Dept. of Electrical and Computer Engineering, Virginia Tech
What: Presentation - Design Automation Meets Artificial Intelligence
Where: Room 246 Nichols Hall, KU West Campus
When: 2:00 p.m. Monday, May 1
Why: Refreshments will be served

Abstract: Electronic Design Automation (EDA) has been around for more than half a century. They have brought us tools that range from analyzing circuits to performing circuit optimization, verification and synthesis. In all the EDA tasks, formal methods have played a major role in realizing the automation goals. Artificial intelligence (AI), on the other hand, have played limited roles. With the recent advances in AI, the potential of EDA can be seen in a new light. Not only can it realize what humans can do, it can probably and most likely also accomplish what humans cannot, such as automatically designing bug-free, functionally correct hardware and software from a given spec. In this talk, we will discuss the advances made in functional verification and automated synthesis and where the combination of EDA and AI might take us in the coming years to handle and manage larger and more complex systems.

Bio: Michael Hsiao is a professor in the Department of Electrical and Computer Engineering at Virginia Tech. He received the B.S. degree in computer engineering (highest honors) and the M.S. and Ph.D. degrees in electrical engineering from the University of Illinois at Urbana-Champaign in 1992, 1993 and 1997, respectively. His research has been recognized for the most influential papers in the first ten years (1998-2007) of Design Automation and Test Conference in Europe (DATE), best paper award at the 2010 IEEE Asian Test Symposium, the best student paper award at the 2012 IEEE International Test Conference, and the NSF CAREER Award. Over the years, he has served as associate editor on IEEE Trans. Computers, ACM Trans. Design Automation of Electronic Systems, as well as on editorial boards of several journals and on program committee for numerous international conferences. His current research interests include design automation of hardware and software. He is a Fellow of IEEE.

Danger, Will Robinson!

2017-03-27T11:33:00.003-07:00

We have all seen the recent headlines that highlight the potential impact of artificial intelligence and autonomous systems on our collective futures—“What jobs will robots take first?”—“Cars that are designed to kill”—“The Internet of Battlefield Things, Robots at War”—“One in three jobs will be eliminated by AI” “Should AI be programmed to feel pain?” Funding agencies are starting to shift from basic AI technology to seeking understanding of social implications and integrated human-machine systems. This topic touches all facets of science and society and is ripe for interdisciplinary exploration. How can we prepare ourselves and our students for the future?

This blog provides a forum for the KU research community to share both current and speculative perspectives on artificial intelligence, automation, and robotics.