- Ai - Engineering Mailing Lists

Meeting time for AI alignment reading group
by Pope, Quintin 13 Nov '21

13 Nov '21

Hello everyone, The survey results are in. Our first meeting will be at 2 PM (PST) on Friday, November 19th, and will repeat weekly. For our first discussion, I thought we'd start with something short and optimistic: Alignment by Default <https://www.lesswrong.com/posts/Nwgdq6kHke5LY692J/alignment-by-default>. It argues that powerful models may have internal "natural abstractions" that straightforwardly represent human values. Some combination of finetuning, interpretability tools and luck may be enough to "wire up" the human values representation to the model's output and thereby get an aligned model. I'm interested in how plausible alignment by default seems to everyone, whether there are any architecture or training modifications we can make that raise the odds of alignment by default, and in how to best manage the "wire up" step where we get a globally aligned model out of a model with a human values subcomponent. Anyone who's at all interested is welcome to join! Join Zoom Meeting https://oregonstate.zoom.us/j/95843260079?pwd=TzZTN0xPaFZrazRGTElud0J1cnJLU… Password: 961594 Phone Dial-In Information +1 971 247 1195 US (Portland) +1 253 215 8782 US (Tacoma) +1 301 715 8592 US (Washington DC) Meeting ID: 958 4326 0079

1 0

[CoRIS-List] OSU Robotics Seminar, Charlie Kemp and Aaron Edsinger and Friday November 12
by Ross L. Hatton 12 Nov '21

12 Nov '21

Hi All, I'm pleased to announce that this week we will have an OSU Robotics seminar from Aaron Edsinger and Charlie Kemp of the mobile-manipulation company Hello Robot. (Dr. Kemp is also a professor at Georgia Tech). The seminar will be at 10am Pacific on Friday, November 12. ---- Drs. Edsinger and Kemp will be joining us remotely via Zoom, at https://oregonstate.zoom.us/j/93958424124?pwd=bVlhZlRCbTFOaHIwZUx1R1h6d205Z… ---- If you would like to attend the seminar in person with others at OSU, I will have LINC 302 (the seminar room for this term) set up for Prof. Koppal's presentation, using the new classroom Zoom capabilities. If you would prefer not to attend in person, but would like some unstructured interaction time with others in the OSU Robotics community, I encourage you to join the seminar's Gather space before or after the seminar, https://gather.town/app/HeC6WAB1zL6cKWp5/OSU%20Robotics%20Seminar =================== Title Hello Robot: Democratizing Mobile Manipulation Abstract Mobile manipulators have the potential to improve life for everyone, yet adoption of this emerging technology has been limited. To encourage an inclusive future, Hello Robot developed the Stretch RE1, a compact and lightweight mobile manipulator for research that achieves a new level of affordability. The Stretch RE1 and Hello Robot’s open approach are inspiring a community of researchers to explore the future of mobile manipulation. In this talk, we will present the Stretch RE1, its academic origins, and the growing community and ecosystem around it. We will also touch on connections with Oregon State University, including participation in our first summer internship program, and Hello Robot’s role in the new NSF AI Institute, AI-CARING. We will conclude with a live demo and a question and answer session. Bios Aaron Edsinger, CEO and Cofounder Aaron has a passion for building robots and robot companies. He has founded four companies focused on commercializing human collaborative robots. Two of these companies, Meka Robotics and Redwood Robotics, were acquired by Google in 2013. As Robotics Director at Google, Aaron led the business, product, and technical development of two of Google’s central investments in robotics. Aaron received his Ph.D. from MIT CSAIL in 2007. Charlie Kemp, CTO and Cofounder Charlie (https://charliekemp.com) is a recognized expert on mobile manipulation. In 2007, he founded the Healthcare Robotics Lab at Georgia Tech, where he is an associate professor in the Department of Biomedical Engineering. His lab has conducted extensive research on mobile manipulators to assist older adults and people with disabilities. Charlie earned a B.S., M.Eng., and Ph.D. from MIT. He first met Aaron while conducting research in the MIT Computer Science and Artificial Intelligence Laboratory (CSAIL). =========== Ross L. Hatton Associate Professor, Robotics and Mechanical Engineering Collaborative Robotics and Intelligent Systems Institute Oregon State University coris.oregonstate.edu research.engr.oregonstate.edu/lram/ rosslhatton.com ross.hatton(a)oregonstate.edu

2 1

AI seminar: Nov 10 2021
by Mangannavar, Rajesh Devaraddi 10 Nov '21

10 Nov '21

Dear all, Our next AI seminar on *"**Estimating Long-term Rewards by Off-policy Reinforcement Learning" *by Lihong Li is scheduled to be on November 10th (Tomorrow), 1-2 PM PST. There will be *NO *graduate student Q&A session. Hence everyone is encouraged to ask questions during the seminar. Note that this is a zoom event. Zoom Link: https://oregonstate.zoom.us/j/93591935144?pwd=YjZaSjBYS0NmNUtjQzBEdzhPeDZ5U… *Estimating Long-term Rewards by Off-policy Reinforcement Learning * Lihong Li Senior Principal Scientist Amazon *Abstract*: One of the core problems in reinforcement learning (RL) is estimating the long-term reward of a given policy. In many real-world applications such as healthcare, robotics and dialogue systems, running a new policy on users or robots can be costly or risky. This gives rise to the need for off-policy, or counterfactual, estimation: estimate the long-term reward of a given policy using data previously collected by another policy (e.g., the one currently deployed). This talk will describe some recent advances in this problem, for which many standard estimators suffer an exponentially large variance (known as "the curse of horizon"). Our approach is based on a dual linear program formulation of the long-term reward, and can be extended to estimate confidence intervals. *Bio*: Lihong Li is a Senior Principal Scientist at Amazon. He obtained a PhD degree in Computer Science from Rutgers University. After that, he held research positions in Yahoo!, Microsoft and Google, before joining Amazon. His main research interests are in reinforcement learning, including contextual bandits, and other related problems in AI. His work is often inspired by applications in recommendation, advertising, Web search and conversational systems. Homepage: http://lihongli.github.io *Please watch this space for future AI Seminars :* * https://eecs.oregonstate.edu/ai-events <https://eecs.oregonstate.edu/ai-events>* Rajesh Mangannavar, Graduate Student Oregon State University ---- AI Seminar Important Reminders: -> The AI Seminar has a strict "no electronics" and "no recordings" policy. -> For graduate students in the AI program, attendance is strongly encouraged.

2 1

AI seminar: Nov 10 2021
by Mangannavar, Rajesh Devaraddi 05 Nov '21

05 Nov '21

Dear all, Our next AI seminar on *"* *Estimating Long-term Rewards by Off-policy Reinforcement Learning" *by Lihong Li is scheduled to be on November 10th, 1-2 PM PST. Note that this is a zoom event. Zoom Link: https://oregonstate.zoom.us/j/93591935144?pwd=YjZaSjBYS0NmNUtjQzBEdzhPeDZ5U… *Estimating Long-term Rewards by Off-policy Reinforcement Learning * Lihong Li Senior Principal Scientist Amazon *Abstract*: One of the core problems in reinforcement learning (RL) is estimating the long-term reward of a given policy. In many real-world applications such as healthcare, robotics and dialogue systems, running a new policy on users or robots can be costly or risky. This gives rise to the need for off-policy, or counterfactual, estimation: estimate the long-term reward of a given policy using data previously collected by another policy (e.g., the one currently deployed). This talk will describe some recent advances in this problem, for which many standard estimators suffer an exponentially large variance (known as "the curse of horizon"). Our approach is based on a dual linear program formulation of the long-term reward, and can be extended to estimate confidence intervals. *Bio*: Lihong Li is a Senior Principal Scientist at Amazon. He obtained a PhD degree in Computer Science from Rutgers University. After that, he held research positions in Yahoo!, Microsoft and Google, before joining Amazon. His main research interests are in reinforcement learning, including contextual bandits, and other related problems in AI. His work is often inspired by applications in recommendation, advertising, Web search and conversational systems. Homepage: http://lihongli.github.io *Please watch this space for future AI Seminars :* * https://eecs.oregonstate.edu/ai-events <https://eecs.oregonstate.edu/ai-events>* Rajesh Mangannavar, Graduate Student Oregon State University ---- AI Seminar Important Reminders: -> The AI Seminar has a strict "no electronics" and "no recordings" policy. -> For graduate students in the AI program, attendance is strongly encouraged.

1 0

Meeting time for AI Alignment Reading group
by Pope, Quintin 04 Nov '21

04 Nov '21

Hello everyone, I'm looking to find a weekly timeslot for the AI alignment reading group that works for everyone. If you're interested, please fill out this whenisgood survey <http://whenisgood.net/nszizrb> with your availability. The timezone is UTC-07 (Corvallis time). To clarify, the indicated times represent the start of the one hour reading group meetings. All the best, Quintin (I forgot to include @ai in my original message. Apologies to everyone on the alignment mailing list who got this twice!)

1 0

seminar is about to start - be there
by Fern, Alan Paul 03 Nov '21

03 Nov '21

https://oregonstate.zoom.us/j/93591935144?pwd=YjZaSjBYS0NmNUtjQzBEdzhPeDZ5U…

1 0

AI seminar: Nov 3 2021
by Mangannavar, Rajesh Devaraddi 02 Nov '21

02 Nov '21

Dear all, Our next AI seminar on *"Modularity and Compositionality in Multi-Step Robot Manipulation" *by Caelan Garrett is scheduled to be on November 3rd (tomorrow), 1-2 PM PST. It will be followed by a 30 minute Q&A session by the graduate students. Note that this is a zoom event. Zoom Link: https://oregonstate.zoom.us/j/93591935144?pwd=YjZaSjBYS0NmNUtjQzBEdzhPeDZ5U… *Modularity and Compositionality in Multi-Step Robot Manipulation* Caelan Garrett Research Scientist NVIDIA *Abstract:*We seek to program a robot to autonomously complete complex tasks in a variety of real-world settings involving different environments, objects, manipulation skills, degrees of observability, initial states, and goal objectives. In order to successfully generalize across these settings, we take a model-based approach to building the robot's policy, which enables it to reason about the effects of it executing different sequences of parameterized manipulation skills. Specifically, we introduce a general-purpose hybrid planning framework that uses streams, modules that encode sampling procedures, to generate continuous parameter-value candidates. We present several domain-independent algorithms that efficiently combine streams in order to solve for parameter values that jointly satisfy the constraints necessary for a sequence of skills to achieve the goal. Each stream can be either engineered to perform a standard robotics subroutine, like inverse kinematics and collision checking, or learned from data to capture difficult-to-model behaviours, such as pouring, scooping, and grasping. Streams are also able to represent probabilistic inference operations, which enables our framework to plan in belief space and intentionally select actions that reduce the robot's uncertainty about the unknown world. We demonstrate the generality of our approach by applying it to several real-world tabletop, kitchen, and construction tasks and show that it can even be effective in settings involving objects that the robot has never seen before. *Speaker Bio:* Caelan Garrett is a research scientist at NVIDIA's Seattle Robotics Lab which is led by Professor Dieter Fox. He received his PhD at MIT in the Learning and Intelligent Systems group within CSAIL where he was advised by Professors Tomás Lozano-Pérez and Leslie Pack Kaelbling. His research is on integrating robot motion planning, discrete AI planning, and machine learning to flexibly and efficiently plan for autonomous mobile manipulators operating in human environments. He recently authored the first survey paper on integrated task and motion planning. He is a recipient of the NSF Graduate Research Fellowship. He has previously interned in the autonomous vehicle industry while at Optimus Ride and in the autonomous fulfilment industry while at Amazon Robotics. This talk will be followed by a talk on the topic "*Estimating Long-term Rewards by Off-policy Reinforcement Learning*" by Lihong Li on 10th November 2021. *Please watch this space for future AI Seminars: * * https://eecs.oregonstate.edu/ai-events <https://eecs.oregonstate.edu/ai-events>* Rajesh Mangannavar, Graduate Student Oregon State University ---- AI Seminar Important Reminders: -> The AI Seminar has a strict "no electronics" and "no recordings" policy. -> For graduate students in the AI program, attendance is strongly encouraged.

1 0

AI seminar: Nov 3 2021
by Mangannavar, Rajesh Devaraddi 29 Oct '21

29 Oct '21

Dear all, Our next AI seminar on *"Modularity and Compositionality in Multi-Step Robot Manipulation" *by Caelan Garrett is scheduled to be on November 3rd, 1-2 PM PST. It will be followed by a 30 minute Q&A session by the graduate students. Note that this is a zoom event. Zoom Link: https://oregonstate.zoom.us/j/93591935144?pwd=YjZaSjBYS0NmNUtjQzBEdzhPeDZ5U… *Modularity and Compositionality in Multi-Step Robot Manipulation* Caelan Garrett Research Scientist NVIDIA *Abstract:*We seek to program a robot to autonomously complete complex tasks in a variety of real-world settings involving different environments, objects, manipulation skills, degrees of observability, initial states, and goal objectives. In order to successfully generalize across these settings, we take a model-based approach to building the robot's policy, which enables it to reason about the effects of it executing different sequences of parameterized manipulation skills. Specifically, we introduce a general-purpose hybrid planning framework that uses streams, modules that encode sampling procedures, to generate continuous parameter-value candidates. We present several domain-independent algorithms that efficiently combine streams in order to solve for parameter values that jointly satisfy the constraints necessary for a sequence of skills to achieve the goal. Each stream can be either engineered to perform a standard robotics subroutine, like inverse kinematics and collision checking, or learned from data to capture difficult-to-model behaviours, such as pouring, scooping, and grasping. Streams are also able to represent probabilistic inference operations, which enables our framework to plan in belief space and intentionally select actions that reduce the robot's uncertainty about the unknown world. We demonstrate the generality of our approach by applying it to several real-world tabletop, kitchen, and construction tasks and show that it can even be effective in settings involving objects that the robot has never seen before. *Speaker Bio:* Caelan Garrett is a research scientist at NVIDIA's Seattle Robotics Lab which is led by Professor Dieter Fox. He received his PhD at MIT in the Learning and Intelligent Systems group within CSAIL where he was advised by Professors Tomás Lozano-Pérez and Leslie Pack Kaelbling. His research is on integrating robot motion planning, discrete AI planning, and machine learning to flexibly and efficiently plan for autonomous mobile manipulators operating in human environments. He recently authored the first survey paper on integrated task and motion planning. He is a recipient of the NSF Graduate Research Fellowship. He has previously interned in the autonomous vehicle industry while at Optimus Ride and in the autonomous fulfilment industry while at Amazon Robotics. This talk will be followed by a talk on the topic "*Estimating Long-term Rewards by Off-policy Reinforcement Learning*" by Lihong Li on 10th November 2021. *Please watch this space for future AI Seminars: * * https://eecs.oregonstate.edu/ai-events <https://eecs.oregonstate.edu/ai-events>* Rajesh Mangannavar, Graduate Student Oregon State University ---- AI Seminar Important Reminders: -> The AI Seminar has a strict "no electronics" and "no recordings" policy. -> For graduate students in the AI program, attendance is strongly encouraged.

1 0

[CoRIS-List] Jay Patravali -- Masters Defense: Oct 26th, 12PM Pacific | KEC 1007 and Virtual Zoom Meeting
by Patravali, Jay Sanjay 25 Oct '21

25 Oct '21

Dear all, I'll be defending my thesis on Tuesday 26th October at 12:00 PM in KEC 1007 and virtually through zoom. Feel free to join either way if you're interested! *Title:* Annotation-Efficient Video Representation Learning *Date & Time:* Tuesday, October 26th at 12:00 pm *Place:* KEC1007 and Zoom Meeting Room. *Abstract:* Labeling videos is costly, time-consuming, and tedious. These costs can escalate in applications such as medical diagnosis, etc. where we need domain expertise for annotation. Few-shot learning in Videos aims to solve this problem by building annotation-efficient learning mechanisms. This thesis presents MetaUVFS as the first Unsupervised Meta-learning algorithm for Video Few-Shot action recognition. MetaUVFS leverages over 550K unlabeled videos to train a two-stream 2D and 3D CNN architecture via contrastive learning to capture the appearance-specific spatial and action-specific Spatio-temporal video features respectively. MetaUVFS comprises a novel Action-Appearance Aligned Meta-adaptation~(A3M) module that learns to focus on the action-oriented video features in relation to the appearance features via explicit few-shot episodic meta-learning over unsupervised hard-mined episodes. Our action-appearance alignment and explicit few-shot learner conditions the unsupervised training to mimic the downstream few-shot task, enabling MetaUVFS to significantly outperform all state-of-the-art unsupervised methods on few-shot benchmarks. Moreover, unlike previous few-shot action recognition methods that are supervised, MetaUVFS needs neither base-class labels nor a supervised pre-trained backbone. Thus, we need to train MetaUVFS just once to perform competitively or sometimes even outperform state-of-the-art supervised methods on popular HMDB51, UCF101, and Kinetics100 few-shot datasets. *Advisor:* Dr. Fuxin Li *Committee:* Dr. Fuxin Li, Dr. Sinisa Todorovic, Dr. Stefan Lee, Dr. Yelda Turkan *Zoom Meeting:* Zoom Meeting details: Paravali, Jay Sanjay is inviting you to a scheduled Zoom meeting. Topic: Jay Patravali MS Robotics Thesis Defense. Time: Oct 26, 2021 12:00 PM Pacific Time (US and Canada) Join Zoom Meeting https://oregonstate.zoom.us/j/91663115146?pwd=bTBXQmFZa1VxM2N6emhxamhEenVrZ… Password: 943745 Phone Dial-In Information +1 971 247 1195 US (Portland) +1 253 215 8782 US (Tacoma) +1 301 715 8592 US (Washington DC) Meeting ID: 916 6311 5146 Join by Polycom/Cisco/Other Room System 91663115146(a)zoomcrc.com *EECS Department Response on serving Snacks or Coffee:* "The current OSU policy does not allow for serving of such items for events unless they are individually catered. Occupants would need to supply their own coffee and cookies." Thanks, Jay Patravali

1 0

AI Seminar postponed
by Prasad Tadepalli 25 Oct '21

25 Oct '21

All, Due to some scheduling conflicts, this week's AI seminar on 10/27 by Lihong Li is postponed to November 10. There will not be a seminar this week, but there will be a seminar by Caelan Garrett on November 3 which will be posted at https://eecs.oregonstate.edu/ai-events Prasad

1 0