January 2022 - Ai - Engineering Mailing Lists

Alignment Reading Group Feedback
by Pope, Quintin 12 Jan '22

12 Jan '22

Hello everyone, Attendance for recent reading groups has dropped off. I've been wondering if there's an issue with the timeslot or content. Please let me know if you have any conflicts with the current time (Friday at 2 PM PST). Currently, reading group content focuses strongly on interpretability. If you'd be more interested in some other aspect of alignment research, please let me know. For this week, we'll meet at 2 PM PST on Friday. We'll be discussing "A Mathematical Framework for Transformer Circuits <https://transformer-circuits.pub/2021/framework/index.html>" In this paper, we attempt to take initial, very preliminary steps towards > reverse-engineering transformers. Given the incredible complexity and size > of modern language models, we have found it most fruitful to start with the > simplest possible models and work our way up from there. Our aim is to > discover simple algorithmic patterns, motifs, or frameworks that can > subsequently be applied to larger and more complex models. Specifically, > in this paper we will study transformers with two layers or less which > have only attention blocks – this is in contrast to a large, modern > transformer like GPT-3, which has 96 layers and alternates attention blocks > with MLP blocks. > 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 All the best, Quintin

1 0

[CoRIS-List] OSU Robotics Seminar, Friday January 28: Sandhya Saisubramanian
by Ross L. Hatton 11 Jan '22

11 Jan '22

Hi All, We'll be kicking off the winter term of the Robotics Seminar with a talk from Sandhya Saisubramanian, who recently joined the EECS faculty at Oregon State and works in the area of autonomous systems. I'm preparing the a full list of upcoming seminars, and will also send out another announcement of Sandhya's talk closer to the date. -Ross =========== 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 =========== Title: Avoiding Negative Side Effects of Autonomous Systems in the Open World Abstract: Autonomous systems are increasingly deployed in open-world environments. These systems have broad societal impacts and it is critical to ensure that they operate safely when deployed. Due to the practical challenges in data collection and precise model specification, AI systems often operate based on imperfect models which may lead to undesirable behaviors. A particular form of unexpected, undesirable consequence is the negative side effects. In this talk, I will (1) discuss the challenges in identifying and mitigating the impacts of negative side effects; (2) examine the efficiency of different forms of feedback to learn the penalty associated with negative side effects; and (3) discuss a multi-objective approach to mitigate the side effects. Bio: Sandhya Saisubramanian is an Assistant Professor in the School of EECS at Oregon State University. She recently completed her Ph.D. from the University of Massachusetts Amherst. Her research interests are in developing general techniques for reliable decision-making in autonomous systems that operate in fully and partially observable open-world environments. She is a recipient of Distinguished Paper Award at IJCAI 2020 and was selected as a Rising Star in EECS by UC Berkeley in 2020.

1 0

AI Seminar: Hannah Kerner
by Tadepalli, Prasad 11 Jan '22

11 Jan '22

Dear all, The rest of the AI Seminars this term are going to be over Zoom only. This week's seminar speaker, Hannah Kerner of University of Maryland, has been nominated by graduate students. The seminar will be followed by a Q&A session with students. Please note that the recordings of most of the past seminars are available from the AI Seminar page: https://eecs.oregonstate.edu/ai-events See you at the seminar tomorrow. Prasad Tadepalli -------------------- AI Seminar: Advancing Global Food Security and Sustainable Development with Machine Learning and Earth Observations<https://eecs.oregonstate.edu/colloquium/ai-seminar-advancing-global-food-se…> Hannah Kerner Assistant Research Professor Department of Geographical Sciences University of Maryland Wednesday, January 12, 2022 - 1:00pm to 2:00pm https://oregonstate.zoom.us/j/93591935144?pwd=YjZaSjBYS0NmNUtjQzBEdzhPeDZ5U… Abstract Satellite Earth observation (EO) data is rapidly gaining interest in the AI community due to the massive datasets involved as well as the opportunities for using AI and EO data to address urgent challenges related to climate change, the environment, agriculture and food security, and humanitarian needs. However there are currently many challenges for developing AI systems that use EO data for practical applications, namely, there are limited public labeled datasets, a lack of harmonization across labels and source data, and existing algorithms do not account for geographic context. In this talk, I'll present some of the approaches we are developing to address these challenges to create AI+EO systems that can be integrated into real-world applications for advancing global food security and other sustainable development goals. Speaker Bio Dr. Hannah Kerner is an Assistant Research Professor in the Department of Geographical Sciences at the University of Maryland. Her research focuses on developing new machine learning methods for geospatial and remote sensing data and applications including agricultural monitoring, food security, Earth science, and planetary science. Dr. Kerner is the Machine Learning Lead and US Co-Lead for NASA Harvest, NASA's agriculture and food security initiative run out of the University of Maryland.

1 0

Alan Fern’s ai seminar zoom only
by Tadepalli, Prasad 05 Jan '22

05 Jan '22

Please note that today’s AI seminar will be on zoom only. AI Seminar: How to Train Your Robot to Stand, Walk, Run, Hop, and Skip Speaker: Alan Fern Wednesday, January 5, 2022 - 1:00pm to 2:00pm PST Zoom link: https://oregonstate.zoom.us/j/93591935144?pwd=YjZaSjBYS0NmNUtjQzBEdzhPeDZ5U… More details at https://eecs.oregonstate.edu/ai-events

1 0

AI Alignment Reading Group
by Pope, Quintin 04 Jan '22

04 Jan '22

Hello everyone, Our next meeting will be this Friday at 2 PM, PST. We're continuing to explore transformer interpretability. Our next paper is "Knowledge Neurons in Pretrained Transformers <https://arxiv.org/abs/2104.08696>" (GitHub implementation <https://github.com/EleutherAI/knowledge-neurons>). This paper is able to identify where pretrained transformers store knowledge and suppress, amplify or modify that knowledge. If you're at all interested in how transformers learn and represent information, you're welcome to attend! Abstract: > Large-scale pretrained language models are surprisingly good at recalling > factual knowledge presented in the training corpus. In this paper, we > explore how implicit knowledge is stored in pretrained Transformers by > introducing the concept of knowledge neurons. Given a relational fact, we > propose a knowledge attribution method to identify the neurons that express > the fact. We present that the activation of such knowledge neurons is > highly correlated to the expression of their corresponding facts. In > addition, even without fine-tuning, we can leverage knowledge neurons to > explicitly edit (such as update, and erase) specific factual knowledge for > pretrained Transformers. > 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 All the best, Quintin

1 0