The Austin Forum on Technology & Society
  • About
  • Partners
  • Events
  • Podcast
  • Blog
  • Slack
  • News
  • Give
  • Contact

NEWS 

Get your event recaps, news, and partner contributed content. Check out our podcasts and youtube.
Join the mailing list to stay informed!

AI: The Power, Promise & Pragmatic Potential

6/14/2019

0 Comments

 
Picture
Here’s what Happened at the Austin Forum in June
​

Thank you to everyone who attended our event at the Central Austin Library and to our speakers Ganesh Padmanabhan of (Molecula Corp), Randi Ludwig (Dell Technologies), Usman Shuja (SparkCognition), and Matt Sanchez (CognitiveScale) for sharing their insights and expertise on the power of AI technology.
Congratulations to Colin Van Ert for the question that won him a SXSW Interactive 2019 badge. 
"How do we design AI technologies so it benefits all members of society and not just a privileged few?"
Colin will be enjoying all that SXSW Interactive has to offer in March 2020 including tracks on Blockchain & Cryptocurrency, AR/VR/MR, Cities, Government & Politics, Coding, Intelligent Future, and more. 
 

Join us next month for your chance to win a 2020 SXSW Interactive badge. Find out how at the July event. 

Missed the event? Catch up on these:.
Event Slides
#AIatAF
Event Photos
The event featured four speakers who are experts on AI and who brought complementary expertise and perspectives.
 
Ganesh Padmanabhan, Chief Revenue Office of Molecula, was the first and primary speaker. Ganesh's key points included:
  • AI doesn't always replace human intelligence; more commonly, it augments human capability ('augmented intelligence'). The future is not humans being replaced by computers/AIs, but humans and AIs working together. 
  • There are no general AIs, only specialized ones for specific tasks--but there is a wide range of them designed to mimic or enhance specific human cognitive functions
  • Machine learning, the current popular branch of AI, is different than classic, rules-based software because it involves actual learning from data (training a model), so that the trained model can 'sense,' 'think,' and then 'act'. However, machine learning requires a LOT of data to train a model to make good decisions/actions
  • Data is being produced at an exponentially growing rate. Coupled with rapid increases in processor technologies resulting in a declining cost of computing, plus maturing algorithmic techniques, machine learning and deep learning--a very sophisticated subset of machine learning--are not feasible and practical
  • AI is advancing and being adopted so rapidly that the expected impact on the GDP is $15.7 Trillion (!) by 2030
  • Yes, the power of AI will cause some jobs to be replaced/eliminated--but as with every new technology, it will also create many new jobs, including many that don't even exist today!
  • Companies want AI to achieve competitive advantages and drive new innovation. In the US, the private sector is contributing significantly to advances in AI

Randi Ludwig, a data scientist at Dell Technologies, followed Ganesh with a brief overview of data science.  
  • Analytics and statistics are important, but are not data science 
  • Data science uses the scientific method to develop a data-enabled model that can help solve business problems. It still starts with a business problem, from which a data science problem is formulated, then a data science model is developed, and that contributes to a business solution
  • There are some problems that can be solved by simple rules, while other require advanced analytics--and more complex problems may benefit from a data science approach. For example, a simple rules model for predicting rain might be 'are there clouds?' but this would not be very accurate for something as complex as weather. An analytics approach might give a percentage change based on cloud cover. A data science approach would use much more data--types of clouds, humidity, temperature, winds, pressure, etc.--and be still more accurate.

Usman Shuja of SparkCognition spoke next on industrial AI.  
  • He showed examples such as autopilot systems based on machine learning models (trained by data) that are able to infer and make decisions better than previous generations of autopilot models.
  • Predictive maintenance is another important industrial application. For example, offshore oil platforms are very complex systems, and preventing failures saves huge amounts of money and potentially prevents major environmental impacts as well. 
  • In the future, AI will help automate plant/factory operations, with maximum efficiency, productivity and reliability--and safety!
  • Military applications are also important for national defense and security. AI can not only help weapons be more accurate but can more benign impacts like improving efficiency and reliability of equipment. 
​
Finally, Matt Sanchez of CognitiveScale closed with a discussion of ethics and fairness in AI
  • It is important to ensure that AIs are fair and robust, not just accurate and fast  .
  • Automated decisions like retail need to be fair and unbiased, but this is even more true for banking decisions (ensuring no discrimination on loan approvals, for example) or healthcare decisions (e.g who gets which treatment)
  • We need to develop ethics for human-machine interactions.
  • While we need to use lots of data for machine learning approaches in AI, it is important to use the right data (which may not be all of the data) to determine the right actions (effective) without bias
  • There are lots of questions in ethics, fairness, and transparency in AI to be answered. If we don't address them, we could experience a backlash and regulations against AI.
  • Thus, we need to take an ethics based approach for responsible AI. This means considering ALL of the stakeholders and making sure they are all treated fairly. They all have something to say about whether it’s fair and responsible.

Picture
Picture
Special thanks to Trifecta on 3rd for another great after-party! Our attendees have shared they enjoy keeping the conversations going after our presentations and networking receptions, so keep an eye on our website for info on future after-parties with food, drinks, and connection.
Join us on July 9, 2019 for Robotics: Transformational Tools for Industry, Mobility & More ​​with Dr. Mitchell Pryor of The University of Texas Austin and Jared Carl of Nvidia.
​

Tuesday, July 9, 2019

Austin Central Library
710 West Cesar Chavez Street
Austin, TX 78701

5:15pm — Check-in and refreshments


5:45 pm – Doors open for seating

6:15 pm – Presentation
7:30 pm – Networking
​8:15pm - Post-event networking at Trifecta on 3rd



We welcome your participation! Please send us your questions, answers, and prognostications in advance.
Admission to the Austin Forum is always free.​

HELP US CLOSE AUSTIN’S DIGITAL DIVIDE
​
The Austin Forum accepts donations of used smart phones and tablets at all our events.  Donated devices get a factory reset and are set up as new by the team at Austin Pathways’ nationally-recognized “Unlocking the Connection” initiative, which connects every public housing resident with a digital device, digital literacy, and a free or very low-cost internet connection. Your donated smart phone can change lives and help close Austin’s digital divide, thanks to Austin Pathways.​
0 Comments

Your comment will be posted after it is approved.


Leave a Reply.

PARTNERS

Picture
Picture
Picture
Picture
Picture
Picture
Picture
Picture
Picture
Picture
Picture
  • About
  • Partners
  • Events
  • Podcast
  • Blog
  • Slack
  • News
  • Give
  • Contact