Select a calendar:
Filter August Events by Event Type:
Events for August 31, 2023
Thu, Aug 31, 2023 @ 11:00 AM - 01:00 PM
Thomas Lord Department of Computer Science
PhD Dissertation Defense - Baskin B. Senbaslar
Committee Members: Gaurav S. Sukhatme (Chair), Sven Koenig, Satish Kumar Thittamaranahalli, Mihailo R. Jovanovic
Title: Decentralized Real Time Trajectory Planning For Multi Robot Navigation in Cluttered Environments
Abstract: Multi robot collision free and deadlock free navigation in cluttered environments with static and dynamic obstacles is a fundamental problem for many real world applications. Dynamic obstacles can additionally be interactive, i.e., changing their behaviors depending on the behaviors of other objects. We focus on decision making algorithms, with a particular emphasis on decentralized real time trajectory planning, to enable multi robot navigation in such environments.
Practicality of the developed approaches is a central focus of ours, such that we design our systems and algorithms under assumptions that can be realized in the real world. Central concerns of our treatment are embracing on board compute, memory, and storage limitations of robotic systems, not relying on communication for safe operation, and explicitly account for communication imperfections, allowing navigation with imperfect a priori knowledge, embracing controller trajectory tracking errors and accounting for them, working with minimal sensing and estimation capabilities, and achieving highly reactive collision avoidance behavior.
We introduce i. two decentralized real time multi robot trajectory planning algorithms to allow static obstacle, interactive dynamic obstacle, and teammate avoidance, ii. a constraint generation, overconstraining, and constraint discarding scheme to ensure inter robot collision avoidance under asynchronous planning that is inherent in decentralized systems, which we use within one of the proposed planners, and iii. a multi robot aware planning and control stack that allows collision free and deadlock free navigation in diverse types of environments, which combines three qualitatively different decision making approaches in a hierarchical manner.
Audiences: Everyone Is Invited
Contact: Melissa Ochoa
NL Seminar - Phishing Emails, Improvised Explosive Devices and Quantum: A Natural Language Understanding Perspective
Thu, Aug 31, 2023 @ 11:00 AM - 12:00 PM
Information Sciences Institute
Conferences, Lectures, & Seminars
Speaker: Mitch Mithun, USC/ISI
Talk Title: Phishing Emails, Improvised Explosive Devices and Quantum: A Natural Language Understanding Perspective
Series: NL Seminar
Meeting hosts only admit guests that they know to the Zoom meeting. Hence, you are highly encouraged to use your USC account to sign into Zoom.
If you are an outside visitor, please inform us at nlg DASH seminar DASH host AT isi DOT edu beforehand so we will be aware of your attendance and let you in.
In this talk Mitch will present 3 projects he worked on in the past year as part of his post doctoral tenure at ISI. In the first project Mitch will explore his findings and discoveries in an effort to answer the question Why do humans still fall prey sometimes to Phishing emails. Specifically, he will talk about the recent collaborative effort between experts in cyber security and natural language processing in exploring several subtle signals typically found in phishing emails which fool humans and or AI models. He will also present a comparative analysis of performance between humans and AI models on these signals, providing insight into the learning ability of both.
In the second part, Mitch will present his work on how to explain and ground the predictions of Large Language Models from a schema curation perspective. Large Language Models are extremely adept at predicting a novel future event or missing events from a given set of events in a complex event. For example, if you ask Chat GPT to predict what are the key events that happen when an Improvised Explosive Device attack occurs, it will start with A person buys huge amount of Ammonium Nitrate as the first event. However how is this result explainable (and verifiable) by human intuition, given that the training data and or the training process of these LLMs are not publicly available?
In the third part, Mitch will present his work on using Quantum Natural Language Processing QNLP in low resource settings. QNLP is a very nascent field which deals with using quantum computers to solve natural language processing problems. QNLP these models are different than neural network-based models, including GPT, because they incorporate compositionality aka grammar fundamentally, while neural network based models rely on learning context through a bag of words approach. He will show that this advantage of QNLP models is more pronounced in few shot learning settings where the data to be trained on is very low.
Biography: Mitch was a post doctoral researcher at ISI where he was working with Marjorie Freedman and Ralph Weischedel in the networking and cyber security division. Mitch, graduated from his PhD from University of Arizona, along with his Masters in Computer Science, before joining ISI as a postdoc. Before that, he worked in the software industry for 10 plus years as a product manager in a research lab. He also has a master degree in Physics from Birla Institute of Technology and Science BITS, Pilani, India. His research interests include natural language processing, cyber security and quantum computation.
Host: Jon May and Justin Cho
More Info: https://nlg.isi.edu/nl-seminar/
WebCast Link: https://youtu.be/xPrATNWf-8E
Audiences: Everyone Is Invited
Contact: Pete Zamar
Event Link: https://nlg.isi.edu/nl-seminar/