/* ---- Google Analytics Code Below */

Friday, February 23, 2018

Microsoft Expands Bot Framework

Was impressed when I saw a demo of Azure being used to build simple chatbots.  I thought then that it is about what services such systems could work with and their related data.  So why can't these frameworks work with Outlook, Linkedin, Cortana and now Teams?   And other Microsoft capabilities? This looks to be in the right direction.

Microsoft Bringing Python, Java Support to Bot Framework

Posted on February 22, 2018 by Mehedi Hassan

" ... The cross-platform bot-building service currently allows developers build bots that work across services like Skype, Facebook Messenger, Slack, Telegram, and more using JavaScript, and C#.   .... " 

Thursday, February 22, 2018

Blockchain and Artificial Intelligence

Below is from the talk mentioned above from our Linked Cognitive Systems Group.  Join us in future talks, many will be mentioned here.

Blockchain and Artificial Intelligence  by Dr. Vugranam (VC) Sreedhar

Abstract: In this talk I will briefly introduce deep connection between the underlying models of blockchain and artificial intelligence (AI) ..... 

Here are the slides from Dr. Vugranam's Lecture.  Quite technically oriented, but also some general embedded gems.  Following up:  http://cognitive-science.info/wp-content/uploads/2018/02/CSIG_Vugranam_Sreedhar.BlockChain-AI-02-22-2018-v1.pdf

(Update):  And more by Dr Vugranam, this more developer and detail oriented, includes specific information about architecture of Smart Contract blockchain approaches:  https://www.youtube.com/watch?v=qnUBzE9CQqg--

Time Inconsistent Planning

Could such a method be used to plug into process models to include planning functions?

Time-Inconsistent Planning: A Computational Problem in Behavioral Economics  By Jon Kleinberg, Sigal Oren 

Communications of the ACM, Vol. 61 No. 3, Pages 99-107  (Abstract)
10.1145/3176189

In many settings, people exhibit behavior that is inconsistent across time—we allocate a block of time to get work done and then procrastinate, or put effort into a project and then later fail to complete it. An active line of research in behavioral economics and related fields has developed and analyzed models for this type of time-inconsistent behavior.

Here we propose a graph-theoretic model of tasks and goals, in which dependencies among actions are represented by a directed graph, and a time-inconsistent agent constructs a path through this graph. We first show how instances of this path-finding problem on different input graphs can reconstruct a wide range of qualitative phenomena observed in the literature on time-inconsistency, including procrastination, abandonment of long-range tasks, and the benefits of reduced sets of choices. We then explore a set of analyses that quantify over the set of all graphs; among other results, we find that in any graph, there can be only polynomially many distinct forms of time-inconsistent behavior; and any graph in which a time-inconsistent agent incurs significantly more cost than an optimal agent must contain a large "procrastination" structure as a minor. Finally, we use this graph-theoretic model to explore ways in which tasks can be designed to motivate agents to reach designated goals.  .... "

(Full article requires subscription)

Can We Trust a Robot?

From Communnications of the ACM:

How Can We Trust a Robot?" by Benjamin Kuipers, says robots must be designed to understand and follow social norms. Kuipers describes the importance of instilling trust and ethics into robots in an original video at bit.ly/2om4z9X. ... 

" ... "Trust is essential for the successful functioning of society. Trust is necessary for cooperation, which produces the resources society needs. Morality, ethics, and other social norms encourage individuals to act in trustworthy ways, avoiding selfish decisions that exploit vulnerability, violate trust, and discourage cooperation. As we contemplate the design of robots (and other AIs) that perceive the world and select actions to pursue their goals in that world, we must design them to follow the social norms of our society. Doing this does not require them to be true moral agents, capable of genuinely taking responsibility for their actions. .... "


Georgia Tech Tests Assistants in Dorms

Like to see not only what is being done, but what skills are seen as being most useful in varying contexts, here in a Georgia Tech dorm, a university known for practical engineering.  Worked with them on Tech applications.   Can we get a read out about which skills provided end up being most useful? Especially beyond what you would typically call 'hospitality' functions.   Will follow up.

Amazon Alexa pilot begins in Towers dorm  at Ga Tech By Polly Ouellette
On the evening of Feb. 7, all residents of Towers Residence Hall were invited to pick up their very own Amazon Dot, a voice-controlled personal assistant that they would install in their dorm room.

The Dots, which can be paired with a bluetooth speaker, will respond to verbal commands that tell the device to play music, get the weather and listen to the news. Additionally, students will be able to perform several Tech-specific operations.  .... "

(Update) This was previously also tested at AZ State in August 2017.

Augmented Reality Glasses to Transform Vision

The emphasis on vision correction in context is interesting.

DARPA Awards $4.7 Million Grant to Transform Augmented-Reality Glasses 
Military Embedded Systems   By Lisa Daigle

Researchers at Columbia University, Stanford University, and the University of Massachusetts at Amherst have won a $4.7-million U.S. Defense Advanced Research Projects Agency grant to develop a lightweight glass that dynamically monitors the wearer's vision and displays vision-corrected contextual images. The goal is to create an ultra-high-resolution, see-through, head-mounted display with a large field of view and significantly reduced size, weight, and power consumption. The device also will be correct the user's vision deficiencies in real time and project aberration-corrected visible contextual information onto the retina. The augmented reality glass relies on the ultrafast generation of arbitrary wavefronts, both in visible and near-infrared wavelengths. The researchers will develop a scalable fabrication process based on standard complementary metal-oxide semiconductor techniques, and well-established procedures to integrate the new materials into the silicon nitride-integrated photonics platform. The team will develop analytical and computational tools for modeling large resonator arrays and dynamics of device performance.... " 

Protecting Deliveries

More details from the world of direct to home delivery.  A Smart Home then needs to consider the and ensure the security of its acquisitions.

Are smart homes smart enough to foil package thieves?  by Tom Ryan in Retailwire, with further expert commentary

With e-commerce’s rapid expansion, so grow the number of incidents of packages being stolen from consumers’ doorsteps.

According to a survey from Comcast, three in ten Americans who live in houses or townhomes have been victims of package theft. And 53 percent know someone who has had a package stolen from outside their home.

The survey was released timed to the launch of Comcast’s Xfinity Home 24/7 video recording service. The camera is controlled by artificial intelligence to record movement so that a home owner can view suspicious activity outside their home through an app. ...."

Wednesday, February 21, 2018

Amazon, Whole Foods and Prime

Amazon improves connections between their online Prime buying club benefits and their brick and mortar acquisitions.  Whole Food 5% discount with Amazon Rewards Visa.  In the coming weeks I will be testing their 2 hour online direct delivery from a nearby Whole Foods. 

Via FMI Daily Lead:  Amazon will now give Prime members a 5% discount when they shop at Whole Foods Market stores using their Amazon Rewards Visa Cards, the company said. Amazon Prime members pay a $99 annual fee for benefits including free shipping and video streaming, and the Rewards Visa comes with other perks including 5% discounts on Amazon purchases. ... " 

Malicious Use of AI

Non technical, overview look at where we need

AI ripe for exploitation, experts warn   By Jane Wakefield in the BBC
Drones turned into missiles, fake videos manipulating public opinion and automated hacking are just three of the threats from artificial intelligence in the wrong hands, experts have said.

The Malicious Use of Artificial Intelligence report warns that AI is ripe for exploitation by rogue states, criminals and terrorists.

Those designing AI systems need to do more to mitigate possible misuses of their technology, the authors said.

And governments must consider new laws.  ..... 

Predicting Earthquakes

Had seen a number of research efforts here,  another move ...

Today, A.I. helps detect tiny earthquakes. Tomorrow, it might predict the big one

Earthquakes are notoriously difficult to predict. Even major quakes often occur with little warning. Meanwhile, there are many hundreds of thousands of smaller earthquakes that humans rarely ever feel but are occasionally detected on seismographs.

Now, researchers from Harvard University and the Massachusetts Institute of Technology have developed an artificial intelligence (A.I.) neural network to better help detect earthquakes of all sizes. In a recent study published in the journal Science Advances, the A.I. system was shown to be more accurate than current methods, and may help bring seismologists closer to the elusive goal of earthquake prediction. .... " 

Thinking deeply about Reinforcement Learning

Remember reading about reinforcement learning, having designed more mundane neural network deep learning,  and thinking, but how do you design the 'objective function' to drive its action? 

Nicely constrained and goal oriented.  Below, from O'Reilly,  " ... A must-read series of posts by Ben Recht "unpacks what is legitimately interesting and promising in RL and what is probably just hype.".  ... Following.

Make It Happen    nbvcxz By Benjamin Recht  

This is the first part of “An Outsider’s Tour of Reinforcement Learning.” Part 2 is linked to.

If you read hacker news, you’d think that deep reinforcement learning can be used to solve any problem. Deep RL has claimed to achieve superhuman performance on Go, beat atari games, control complex robotic systems, automatically tune deep learning systems, manage queueing in network stacks, and improve energy efficiency in data centers. What a miraculous technology! I personally get suspicious when audacious claims like this are thrown about in press releases, and I get even more suspicious when other researchers call into question their reproducibility. I want to take a few posts to unpack what is legitimately interesting and promising in RL and what is probably just hype. I also want to take this opportunity to argue in favor of more of us working on RL: some of the most important and pressing problems in machine learning are going to require solving exactly the problems RL sets out to solve.  .... " 

Tuesday, February 20, 2018

Spotify to Build Own Speakers?

In CIO: Another competitor in the ranks.  Now from a Music aggregator.  What kind of  other Assistant aspects will it include?  Recall how Spotify was designing music based specifically on consumer demand.

Spotify is working on a line of "category defining" hardware products and is ready to start setting up the manufacturing process.

The streaming music company intends to create a hardware category "akin to Pebble Watch, Amazon Echo, and Snap Spectacles," according to job adverts posted over the past year.

One ad for a senior product manager, posted last April, called for an expert to "define the product requirements for internet connected hardware [and] the software that powers it."

Today, a trio of job adverts (spotted by industry site MusicAlly) have been posted, seeking an "operations manager," "senior project manager: hardware production," and "project manager: hardware production and engineering" for the hardware. The first of those adverts states that "Spotify is on its way [to] creating its first physical products and set-up an operational organization for manufacturing, supply chain, sales and marketing." .... '

Work on 1000 Year Clock Begins

Also known as the most famous part of the Long Now Project.  About long term thinking.  Followed it closely in its early days via our work with the Institute for the Future.

Jeff Bezos writes:

@JeffBezos
Installation has begun—500 ft tall, all mechanical, powered by day/night thermal cycles, synchronized at solar noon, a symbol for long-term thinking—the #10000YearClock is coming together thx to the genius of Danny Hillis, Zander Rose & the whole Clock team! Enjoy the video.

Tweet and Video. 
Clock Details and more video:

" ... The full scale 10,000 Year Clock is now under construction. While there is no completion date scheduled, we do plan to open it to the public once it is ready. The essay below by Long Now board member Kevin Kelly discusses what we hope the Clock will be once complete. This is one of several projects by Long Now to foster long-term thinking in the context of the next 10,000 years.

Clock in the Mountain   by Kevin Kelly

There is a Clock ringing deep inside a mountain. It is a huge Clock, hundreds of feet tall, designed to tick for 10,000 years. Every once in a while the bells of this buried Clock play a melody. Each time the chimes ring, it’s a melody the Clock has never played before. The Clock’s chimes have been programmed to not repeat themselves for 10,000 years. Most times the Clock rings when a visitor has wound it, but the Clock hoards energy from a different source and occasionally it will ring itself when no one is around to hear it. It’s anyone’s guess how many beautiful songs will never be heard over the Clock’s 10 millennial lifespan.

The Clock is real. It is now being built inside a mountain in western Texas. This Clock is the first of many millennial Clocks the designers hope will be built around the world and throughout time. There is a second site for another Clock already purchased at the top of a mountain in eastern Nevada, a site surrounded by a very large grove of 5,000-year-old bristlecone pines. Appropriately, bristlecone pines are among the longest-lived organisms on the planet. The designers of the Clock in Texas expect its chimes will keep ringing twice as long as the oldest 5 millennia-old bristlecone pine. Ten thousand years is about the age of civilization, so a 10K-year Clock would measure out a future of civilization equal to its past. That assumes we are in the middle of whatever journey we are on – an implicit statement of optimism. .... " 

Talk on AI and Blockchain this Thursday

In our CSIG group.  Join us.  Relates to our work underway regarding Smart Contracts ...

Date and Time :  Feb 22, 2018 - 10:30am to 11:00am US Eastern
Zoom meeting Link: https://zoom.us/j/7371462221
Zoom Callin: (415) 762-9988 or (646) 568-7788 Meeting id 7371462221
Zoom International Numbers: https://zoom.us/zoomconference
Website: http://cognitive-science.info/community/weekly-update/

Title: Blockchain and Artificial Intelligence

Abstract: In this talk I will briefly introduce deep connection between the underlying models of blockchain and artificial intelligence (AI).

Bio: Dr. Vugranam (VC) Sreedhar recently joined IBM GTS/TSS from IBM TJ Watson Research Center to lead all aspects of Blockchain solutions for Technical Support Services. He led several successful projects in broad areas, including compilers, programming technology, security, compliance, service delivery and blockchain. He is also ACM (Association of Computing Machinery) Distinguished Scientist and also member of IBM Academy of Technology. He earned is Ph.D. from McGill University, with Dean's Honor.

For a look at upcoming speakers and past presentations visit the CSIG website:   http://cognitive-science.info/community/weekly-update/ 

AI Platforms Extracting Corporate Knowledge

Quite a considerable claim.   Can we extract all the knowledge of a company for easy use? We developed and tried systems with similar goals.  Looking further.

Can This AI Platform Make You A Better Leader?
Half of executives fail within the first 18 months of being promoted or hired. Could software fix the problem?  By Lydia Dishman in FastCompany

A robot might not directly take your job, but chances are that automation will force you to learn new skills. In fact, according to experts at McKinsey, as many as 375 million workers globally may need to switch fields and learn new skills soon. That’s because at least a third of tasks in about 60% of jobs can be automated.

High-level thinking and creativity are beyond the capabilities of artificial intelligence. So how can a software platform claim to make executives and managers better at leading? After all, aren’t qualities like optimism,  empathy, and emotional intelligence rooted firmly in human behavior?

They are, but there are others that plenty of people struggle with like setting and keeping individual and organizational priorities. Enter Indiggo, a platform powered by a proprietary AI tool called “indi” that launched earlier this month.  According to Indiggo’s CEO and cofounder Janeen Gelbart, indi is a brain that has consumed all the knowledge the company has gathered in its 15 years of operation. That’s a massive data set of situational analysis, advice, and guidance for different types of leaders that Gelbart says is quite powerful and can provide a “return on leadership” much like ROI. .... " 

Immersive Terf

Immersive Terf.  Was reminded of Qwaq, which we examined for 3D Immersive collaboration.   To provide a virtual-world style collaboration.   Is it useful to have cartoon-like figures representing you and your collaborating colleagues, to provide an avatar inhabiting world with spaces that represent real conference spaces?  And is this more efficient than using advanced video driven spaces like Cisco's Telepresense? 

Keep Your Customers Coming Back

Good piece.  Not at all technical, but worth repeating what you need to get done.  The tech people I talk to don't seem to know these obvious things, and are always looking for some other tech magic.  But if they don't get these things right, it doesn't matter.

How to Keep Your Customers Coming Back    By Mike Dupuy in CustomerThink

In today’s competitive marketplace, how can you encourage customers to stay loyal to your brand and products when there are so many choices available? You have to provide an outstanding product, of course, and it must be appropriately priced, but there are a lot of companies that can do that. What will set you apart and build brand loyalty, however, is your customer support. .... " 

Monday, February 19, 2018

Samsung Patents the Drone Screen

In our innovation center we looked at many variants of the screen.  Especially as it related to in store and public retail advertising.  Here something new, a drone that resents a screen.  With lots of possibilities around scale.  The current winter Olympics showed how swarms could be impressive, why not displays integrated into such systems.  Nothing demonstrable as yet, just a patent.

Samsung patents a flying screen that could be used for hovering video
It may also be able to adjust its position based on the angle of your gaze.

By Mallory Locklear, @mallorylocklear in Engadget .... " 

GraphGrail AI

New to me, I like the attempt to create a broad solution ecosystem.   Send me some examples of useful applications in place.   Mentions Smart Contracts, which we are currently developing.

GraphGrail:   AI meets Blockchain

Graphgrail AI - is the World's first Artificial Intelligence Platform for Blockchain Built on top of Natural Language Undertstanding Technology with the Dapp's Marketplace. .... 

Decentralized platform, open to the world .... 
Ethereum for data-science professionals ..... 

Whitepaper:  https://en.graphgrail.com/whitepaper/en/index.html




#GraphGrailAi mission: the creation of a strong AI (Artificial general intelligence) that will be open to all, controlled and trained by developers throughout the entire world.... 


Portal for Scientific Discovery

Augmentation for discovery is something we examined for research.

Networking, Data Experts Design a Better Portal for Scientific Discovery
Lawrence Berkeley National Laboratory

Scientists from the U.S. Department of Energy's Energy Sciences Network (ESnet) and the University of Chicago and Argonne National Laboratory's Globus team have designed a new data portal to make information sharing faster, more reliable, and more secure. "Our new design preserves...ease of use, but easily scales up to handle the huge amounts of data associated with today's science," says ESnet's Eli Dart. The new portal design is based on Dart's Science DMZ, a high-performance network framework that connects large-scale data servers directly to high-speed networks and increasingly is used by research institutions for data transfer management. Another key platform is the cloud-based Globus service enabling developers to outsource responsibility for complex tasks such as authentication, authorization, data movement, and data sharing. An important system element is Globus Connect, which lets the Globus service transfer data to and from the computer using high-performance protocols as well as HTTPS for direct access. ....  "

Hacking Phone Pins with Sensor Data

Disturbing Situation ... You have to think about what data is being gathered by sensors, and if that new data can be used to predict other data ... and beyond.  A whole stream of inference to check.  Note the prediction does not necessarily have to be precise, just approximate. 

Hackers can guess your phone's PIN using 'harmless' sensor information
Six key sensors have no permission requirements, leaving their stored data open to any app that wants it .... " 

Dimensions of Digital Trust

Fascinating take.  A little skeptical about the approach to gathering this detailed information.  Why just 4 dimensions?   Worth the read.

The 4 Dimensions of Digital Trust, Charted Across 42 Countries  By Bhaskar Chakravorti, Ajay Bhalla, Ravi Shankar Chaturvedi in the HBR

The year 2018 is barely underway and, already, digital trust initiatives have captured headlines. Facebook’s Mark Zuckerberg has said his platform will de-prioritize third-party publisher content to keep users focused on more “meaningful” posts from family and friends. Google has led off the new year by blocking websites that mask their country of origin from showing up on Google News. And the European Union’s upcoming General Data Protection Regulation (GDPR) will affect every organization around the world that handles personal data for EU residents. The regulations will also, no doubt, inform data protection laws and corporate trust-building strategies elsewhere.

Even China’s opaque behemoths have started the year with unprecedented acknowledgements of the need to address trust concerns: Tencent had to publicly deny that it collects user WeChat history after it was openly challenged; Alibaba’s Ant Financial apologized to users of its mobile-payment service for automatically enrolling them in its social-credit scoring service.

What these stories underscore is that our digital evolution and our productive use of new technologies rests on how well we can build digital trust. But is it possible to measure digital trust and compare it across countries? Are there countries where guaranteeing trust is a more urgent priority and will draw a larger share of trust-building resources and regulations? The Fletcher School at Tufts University and Mastercard have a launched a research initiative to address these questions by studying the state of digital trust across 42 countries. Here are some of our initial findings, drawn from the study, “Digital Planet 2017: How Competitiveness and Trust in Digital Economies Vary Across the World.” ... " 

Drop Ship Stress Panel

Does Drop Ship Put too Much Stress on the Supply Chain?
Dan Gilmore in Retailwire.

Through a special arrangement, presented here for discussion is a summary of a current article from Supply Chain Digest. .... 

A panel at last month’s Retail Value Chain Federation conference in Scottsdale explored the inventory challenges vendors face supporting drop shipping for their retail partners’ online operations.

Since RVCF is a semi-private function, the identities of the three manufacturers and one 3PL on the panel are kept anonymous.

Many approaches exist for managing drop ship inventories, including:

Having a dedicated inventory in a separate location in the DC;

Having one pool of co-located inventory for a company’s own piece pick/e-commerce business plus the retail drop ship;

Separating inventories logically but storing them together, if allocation and warehouse management systems can handle that.   One doesn’t seem to be preferred over the others.

All the panelists, however, tweak inventory availability information to guard against receiving a drop ship order for which they have no inventory, either because of accuracy issues or because someone else grabbed the merchandise first. One vendor reports no inventory to a retailer when the SKU count reaches five or less.

Almost all retailers want inventory information sent daily, although some want it refreshed multiple times throughout the day. The requirements seem to be based on the companies’ level of IT maturity. ... "

Sunday, February 18, 2018

Electronic Skins as Health Sensors

More novel health sensors.

Electronic skin can display a heartbeat on your hand 
You'd know someone's health just by looking at them.

By Jon Fingas, @jonfingas in Engadget.

Electronic skins might not only detect health troubles in the near future, but display them for the world to see. University of Tokyo researchers have developed an e-skin that can measure vital signs like your heartbeat and display them in real time on a skin display. The design blends a breathable nanomesh electrode and stretchable wiring with an array of micro LEDs that can output basic images bending with your body. Others know right away if you need help -- they'd just have to look at your hand (or anywhere else the sensor works) to get an idea of what's wrong. The sensor can pair with a smartphone and transmit its info to the cloud, too. ...   Dai Nippon Printing

(Update)   More in Digital Trends.

Marketing and AI

Artificial Intelligence Interview with ESOMAR  (the WORLD Association for Social, Opinion and Market Research)    By Tom H. C. Anderson  .... 

What Market Researchers Should Know about AI and Machine Learning – A Q&A with ESOMAR Research World on Artificial Intelligence in Marketing Research

ESOMAR: What is your experience with Artificial Intelligence & Machine Learning (AI)? Would you describe yourself as a user of AI or a person with an interest in the matter but with no or limited experience?

TomHCA: I would describe myself as both a user of Artificial Intelligence as well as a person with a strong interest in the matter even though I have limited mathematical/algorithmic experience with AI. However, I have colleagues here at OdinText who have PhD’s in Computer Science and are extremely knowledgeable as they studied AI extensively in school and used it elsewhere before joining us. We continue to evaluate, experiment, and add AI into our application as it makes sense. .... " 

Security Risks for Alexa in Business

Some good points made for the business use of assistants and other attentive systems.  Methods like voice recognition, biometrics  and validation methods will have to be added and tested for such systems to be viable in the enterprise.   This goes beyond trade secrets, and to data transmission of any kind.

A white hat hacker says no company should use Amazon's enterprise version of its virtual assistant if privacy and trade secrets are important to your company.    By Will Yakowicz  ....     in Inc

Saturday, February 17, 2018

Gillette

Gillette: Fat, Happy and Vulnerable in Its Own Fiefdom
By Bob Herbold on his blog .... 

A recent article in Fortune magazine discussed the massive loss of market share by Gillette; moving from 71% when it was acquired by Procter and Gamble in 2005 down to its current 59%. It points out that a key reason for this was that Gillette simply missed the growing consumer interest in an adequate performing, and very reasonably priced, razor. Dollar Shave Club, Harry’s and Schick jumped on this trend. Meanwhile, Gillette simply stuck to its decades-old game plan of evermore sophisticated and complex razors at ever-increasing prices.

How did Gillette become so unaware of reality? Basically, it appears that Gillette was its own isolated fiefdom at Procter and Gamble, basking in prior success. They got away with this because competition was historically weak, and Gillette was making good profits, so P&G management left them alone. ... " 

Cortana Gets IFTTT

Was wondering why this was taking so long.  Crucial not only because it links in data that can be used, but also shows some of the possibilities for future connections.  Now all the major assistant players link to IFTTT.   Have successfully used it for several tests.

Microsoft’s Cortana is finally on IFTTT  in Engadget

Microsoft keeps striving to find Cortana a place in the crowded smart assistant market, and despite losing a minor feature, it's still adding functionality. Today, Cortana added IFTTT, and launched with interactions to link it up with 550 apps and devices. ... " 

Simple Example of TensorFlow

A good, detailed example of using TensorFlow for one of the most well known analytics techniques.  Instructive.

Linear Regression in Tensorflow  by Aaqib Saeed
Predicting house prices in Boston area

Tensorflow is an open source machine learning (ML) library from Google. It has particularly became popular because of the support for Deep Learning. Apart from that it's highly scalable and can run on Android. The documentation is well maintained and several tutorials available for different expertise levels. To learn more about downloading and installing Tensorflow, visit official website.  .... "

Drones in Uncertain Environments

Long been a student of uncertainty.

A little Uncertainty can help Drones dodge Obstacles at High Speeds, says MIT  By James Vincent    @jjvincent in TheVerge

For drones trying to navigate a busy environment like a warehouse or a forest at high speed, the ability to know exactly where they are at all times would seem pretty essential. Not so, say researchers from MIT’s Computer Science and Artificial Intelligence Laboratory (CSAIL), who have a devised a new, efficient way to guide drones around obstacles. The key ingredient? Uncertainty.

With most drones — and, indeed, most self-driving vehicles — navigation starts with a map. To draw one, depth sensors are used to scan the immediate environment which is compiled into a single 3D model. This then tells the vehicle not only where they are at any given moment, but also how to get to their destination. It’s a method commonly known as “simultaneous localization and mapping,” or SLAM.

SLAM has served the community pretty well to date, but it has its downsides. For one, it’s a very intensive process, that needs lots of high-fidelity data and computing power to process it. This is why Waymo and Uber’s recently settled lawsuit was all about LIDAR — the laser-firing sensors used to collect and process depth data. Data is important.

But, this process creates problems at high speeds and with small crafts like drones. They don’t have the time to collect all the data they need, and giving them the processors to understand it all is expensive.     .... "