Select a calendar:
Filter May Events by Event Type:
SUNMONTUEWEDTHUFRISAT
Events for May
-
Open Alpha Game Premiere
Sat, May 02, 2020 @ 06:00 PM - 08:00 PM
Thomas Lord Department of Computer Science
Student Activity
Open Alpha: a cycle of testing a game in an early stage, wherein the test group is much larger than that of closed testing and typically open to anyone who is interested.
Wanna know what we've been up to all semester? Come join us for our final release premiere to take a sneak peak at our FIRST EVER GAME about... window washing?!?! Get an in-depth look at the game's trailer, demo, and a Q&A with our project leads about the game development process and how our first semester went as a club (especially after going all virtual)! If you're interested in getting involved with Open Alpha, this is a great way to see just what kind cool stuff you get to do!
Catch the premiere at bit.ly/openalpha on May 2nd, 6PM PST!
Thank you!
The Open Alpha Team
Location: Online
WebCast Link: bit.ly/openalpha
Audiences: Everyone Is Invited
Contact: Ryan Rozan
-
CANCELED - Computer Science General Faculty Meeting
Wed, May 06, 2020 @ 12:00 PM - 02:00 PM
Thomas Lord Department of Computer Science
Receptions & Special Events
Bi-Weekly regular faculty meeting for invited full-time Computer Science faculty only. Event details emailed directly to attendees.
Location: 101
Audiences: Invited Faculty Only
Contact: Assistant to CS chair
-
PhD Defense - Artem Molchanov
Mon, May 11, 2020 @ 01:30 PM - 03:00 PM
Thomas Lord Department of Computer Science
University Calendar
Ph.D. Defense - Artem Molchanov 5/11 1:30 pm "Data Scarcity in Robotics: Leveraging Structural Priors and Representation Learning"
Ph.D. Candidate: Artem Molchanov
Date: Monday, May 11, 2020
Time: 1:30 PM - 3:00 PM
Committee: Gaurav S. Sukhatme (Chair), Nora Ayanian, Heather Culbertson, Satyandra K. Gupta
Title: Data Scarcity in Robotics: Leveraging Structural Priors and Representation Learning
Abstract:
Recent advances in Artificial Intelligence have benefited significantly from access to large pools of data accompanied in many cases by labels, ground truth values, or perfect demonstrations. In robotics, however, such data are scarce or absent completely. Overcoming this issue is a major barrier to move robots from structured laboratory settings to the unstructured real world. In this thesis, by leveraging structural priors and representation learning, we provide several solutions when data required to operate robotics systems is scarce or absent.
In the first part of this thesis we study sensory feedback scarcity. We show how to use high-dimensional alternative sensory modalities to extract data when primary sensory sources are absent. In a robot grasping setting, we address the problem of contact localization and solve it using multi-modal tactile feedback as the alternative source of information. We leverage multiple tactile modalities provided by piezoresistive and capacitive sensor arrays to structure the problem as spatio-temporal inference. We employ the representational power of neural networks to acquire the complex mapping between tactile sensors and the contact locations. We investigate scarce feedback due to the high cost of measurements. We study this problem in a challenging field robotics setting where multiple severely underactuated aquatic vehicles need to be coordinated. We show how to leverage collaboration among the vehicles and the spatio-temporal smoothness of the ocean currents as a prior to densify feedback about ocean currents to acquire better controllability.
In the second part of this thesis, we investigate scarcity of the data related to the desired task. We develop a method to efficiently leverage simulated dynamics priors to perform sim-to-real transfer of a control policy when no data about the target system is available. We investigate this problem in the scenario of sim-to-real transfer of low-level stabilizing quadrotor control policies. We demonstrate that we can learn robust policies in simulation and transfer them to the real system while acquiring no samples from the real quadrotor. We consider the general problem of learning a model with a very limited number of samples using meta-learned losses. We show how such losses can encode a prior structure about families of tasks to create well-behaved loss landscapes for efficient model optimization. We demonstrate the efficiency of our approach for learning policies and dynamics models in multiple robotics settings.
Meeting Links:
Zoom Meeting:
https://usc.zoom.us/j/98384775690
Meeting ID: 983 8477 5690
Google Meet (ONLY A BACKUP - IF WE EXPERIENCE PROBLEMS WITH ZOOM):
Meeting ID:
meet.google.com/unz-zwzs-nxj
Phone Numbers:
+1 470-735-5928
PIN: 476 191 520#WebCast Link: https://usc.zoom.us/j/98384775690
Audiences: Everyone Is Invited
Contact: Lizsl De Leon
-
USC Games Expo 2020
Tue, May 12, 2020 @ 02:00 PM - 09:30 PM
Thomas Lord Department of Computer Science
Receptions & Special Events
The USC Games Expo, in Partnership with jam City, will be held on May 12th. The event will be streamed LIVE on multiple platforms - including Twitch - and will be accessible via uscgamesexpo.com
RSVP now at http://uscgamesexpo2020.eventbrite.comLocation: Online
WebCast Link: http://uscgamesexpo.com
Audiences: Everyone Is Invited
Contact: Ryan Rozan
-
PhD Defense - Zhiyun Lu
Wed, May 20, 2020 @ 12:00 PM - 02:00 PM
Thomas Lord Department of Computer Science
University Calendar
Ph.D. Candidate: Zhiyun Lu
Date: Wednesday, May 20, 2020
Time: 12:00 PM - 2:00 PM
Committee: Fei Sha (Chair), Haipeng Luo, C.-C. Jay Kuo
Title: Leveraging Training Information for Efficient and Robust Deep Learning
Abstract: Deep neural nets have exhibited great success on a wide range of machine learning problems across various domains, such as speech, image, and text. Despite decent prediction performances, there are rising concerns for the `in-the-lab' machine learning models to be vastly deployed in the wild. In this thesis, we study two of the main challenges in deep learning: efficiency, computational as well as statistical, and robustness. We describe a set of techniques to solve the challenges by utilizing information from the training process intelligently. The solutions go beyond the common recipe of a single point estimate of the optimal model.
The first part of the thesis studies the efficiency challenge. We propose a budgeted hyper-parameter tuning algorithm to improve the computation efficiency of hyper-parameter tuning in deep learning. It can estimate and utilize the trend of training curves to adaptively allocate resources for tuning, which demonstrates improved efficiency over state-of-the-art tuning algorithms. Then we study the statistical efficiency on tasks with limited labeled data. Specifically we focus on the task of speech sentiment analysis. We apply pre-training using automatic speech recognition data, and solve sentiment analysis as a downstream task, which greatly improves the data efficiency of sentiment labels.
The second part of the thesis studies the robustness challenge. Motivated by the resampling method in statistics, we study the uncertainty estimate of neural networks by local perturbative approximations. We propose to sample replicas of the model parameters from a Gaussian distribution to form a pseudo-ensemble. The ensemble predictions are used to estimate the uncertainty of the original model, which improves its robustness against invalid inputs.
Meeting links:
Zoom: https://usc.zoom.us/j/96089712182 (Meeting ID: 960 8971 2182)
Google Meet (backup): meet.google.com/nxz-eybf-urw
WebCast Link: https://usc.zoom.us/j/96089712182
Audiences: Everyone Is Invited
Contact: Lizsl De Leon
-
Fighting COVID-19 with APIs and Big Data hosted by Bit Project of UC-Davis (RSVP in Advance)
Thu, May 21, 2020 @ 06:00 PM - 07:30 PM
Thomas Lord Department of Computer Science
Workshops & Infosessions
Fighting COVID-19 with APIs and Big Data hosted by Bit Project of UC-Davis.
Bit Project would like to invite students from University of Southern California to participate in our webinar about applying the power of APIs and Big Data to combat COVID-19. Utilizing data provided by Johns Hopkins University, we will be guiding students through Postman to exhibit how code can influence issues faced at a global scale. In addition, we have speakers from reputable organizations, such as Microsoft and Postman, that will explain the role of distinct computing features in the age of COVID-19.
By the end of the webinar, attendees will have experience with:
1. Finding and accessing public APIs
2. Importing, cleaning, and processing geospatial data
3. Using Jupyter Notebook
4. Using Facebook's Prophet Library to forecast the spread of the disease.
While this webinar is centered around APIs and Big Data, we will also have speaker(s) from the front-lines of COVID-19 to conceptually understand the potential magnitude technologists may have on this pandemic.
We will be hosting the webinar on Thursday, May 21st at 6:00 p.m.
To sign-up for the webinar, please RSVP via the following link:
https://www.eventbrite.com/e/fighting-covid19-with-apis-and-big-data-tickets-104366926286
Location: Online
WebCast Link: https://www.eventbrite.com/e/fighting-covid19-with-apis-and-big-data-tickets-104366926286
Audiences: Everyone Is Invited
Contact: Ryan Rozan
