Episode 48: The Inner-Machine Therapist

Voices from DARPA - Un pódcast de DARPA

Categorías:

  In this episode of the Voices from DARPA podcast, John-Francis Mergen (https://www.darpa.mil/staff/mr-john-francis-mergen), a program manager since 2020 in the agency’s Information Innovation Office (https://www.darpa.mil/about-us/offices/BTO), recounts how his interest in science took off as a child when he received a gift of a low-power magnifier from a family friend who was a geologist. From that gift, Mergen says, he learned about the power of observation and of the mindset one brings into that elemental component of the scientific enterprise. For his part, Mergen has spent a lot of time observing the complex ebbs and flows of data packets, which are mobile portions of information that race every which way through the internet and then get reassembled on your computer into a web page, a picture, or an email message. One of the first DARPA programs Mergen started to run last year aims to optimize the efficiency of packet traffic and management based on dynamic prioritization of information categories, such as text, voice or images, while preserving privacy and confidentiality for the sender and recipient of those packets. Another program Mergen runs is anticipating emerging threats associated with the exploding population of internet-connected-devices—the Internet-of-Things (IoT)—with an eye on security-enhancing communications protocols. Mergen has skin in the game: he says he has several hundred devices (including an internet-connected beehive!) at home that are connected to the Internet. One of his newest programs, if successful, will deliver technology that applies artificial intelligence to manage IoT devices so that they automatically and securely configure themselves, in his words, “in a way that is useful but not in a way that can be used” by adversaries, criminals, and others seeking to do harm. In a program just getting underway, Mergen envisions vehicles, manufacturing tools, and other technologies with a kind of self-awareness, which would be based on the many sensors, actuation devices, and computers in their designs, along with the ability to leverage this gadget-based self-awareness into automatic adjustments of operations. The payoff? Mergen says it could lead to more capable and longer-lasting technologies that could bring out their own best in changing circumstances. One possibility is that already-deployed technologies would “discover” capabilities they have in specific situations that not even their designers had in mind.  

Visit the podcast's native language site