Machine Question Answering in the Wild
Eunsol Choi | May 9, 2022 | 4:00 - 5:00pm CST | McMurtry Auditorium in Duncan Hall
Recent large-scale language models provide dramatically improved representations of text that we can adapt easily for a wide range of tasks. Encouraged by their success, we now simulate end applications more closely, embracing world knowledge into the task definition. For example, while prior QA benchmarks provided an evidence document that contains a short answer to the question, we now expect models to generate an answer from the question alone. I will discuss new challenges that arise in these more realistic scenarios. First, models need to handle queries where the answer varies based on their temporal or geographical contexts (e.g., who is the vice president?). Second, models need to organize information to craft multi-sentence answers to address complex questions like “why do birds chirp more in the morning?”. Our results suggest that existing models struggle in both dimensions, despite their high automatic lexical match scores. In the final part of the talk, I will present preliminary work and directions for improving QA models in these challenging scenarios.
|
|
2022 Machine Learning Boot Camp: Executives & Beginners
May 19-20, 2022 | 8:00am - 5:00pm CST | BRC at Rice University
This two-day boot camp will provide a broad introduction to Machine Learning (ML) for executives and beginners. Our goal is to give a manager or a CEO in a field such as medicine, finance, manufacturing, or energy an introduction to the essential concepts behind Machine Learning. Additionally, this boot camp is great for non-practicing beginners looking to learn. Taught by a team of Rice professors in the computational sciences, the Machine Learning Boot Camp will integrate lectures describing the basic concepts to state-of-the-art Machine Learning.
The boot camp assumes no specific prerequisites and will highlight a broad set of concepts, ranging from classical supervised (e.g., linear regression, logistic regression and decision trees) and unsupervised methods (e.g., clustering algorithms, principal components analysis) to modern machine learning approaches (e.g., different neural network architectures, min-max games and generative models, deep reinforcement learning), with numerous case studies and applications. Throughout its duration, the workshop will provide practical suggestions to executives with DOs and DON’Ts when using Machine Learning in practice.
Prerequisites: A general level of familiarity with mathematics is assumed, including college-level calculus, though those without this background will still find the boot camp a useful introduction to modern ML.
Click here to see the full 2022 curriculum.
|
|
Announcing the OSG User School 2022
Apply by April 18, 2022 | University of Wisconsin in Madison
If you could access hundreds or even thousands of computers for your scholarly work or the work you support, what could you do? How could it transform your research? What discoveries might be made? Apply by April 18 for the OSG User School 2022, taking place July 25-29, at the beautiful University of Wisconsin in Madison. Participants will learn to use high throughput computing (HTC) to harness vast amounts of computing power for research, applicable to nearly any field of study. Participants at previous schools have come from engineering, life sciences, earth sciences, social sciences, agricultural and animal sciences, economics, medicine, physics, chemistry, and more. Learn more about HTC. Examples of research advanced by HTC.
Through lectures, demonstrations, hands-on exercises, roleplays, and personal consulting with experts in HTC, participants will learn how HTC systems work, how to run and manage many jobs with large datasets and how to get your own work running on available HTC resources.
Ideal candidates are research trainees (e.g. graduate students and post-docs) and/or staff who support researchers with computing work that requires resources beyond a single computer. We also accept faculty, research staff, advanced undergraduates, etc., so tell us about yourself regardless of your current role!
We are planning a completely in-person event and will follow all COVID guidelines that are in effect at that time. The School pays for all basic travel, hotel, and food costs for selected participants. IMPORTANT DATES:
* Application Period (OPEN NOW): 18 March - 18 April 2022
* OSG User School: 25-29 July 2022MORE INFORMATION AND APPLICATIONS:
*Web: https://www.opensciencegrid.org/user-school-2022/
*Email: user-school@opensciencegrid.org
*Twitter: https://twitter.com/OSGUserSchool
Spring 2022 D2K Showcase
April 20, 2022 | 5:00-7:00pm CST | Duncan HallSave the Date for the Spring 2022 D2K Showcase on Wednesday, April 20 at 5pm (Central Time) and celebrate the impact students are making through data science!- Explore D2K students’ end-of-semester projects covering a wide range of data science and AI topics including computer vision, natural language processing, deep learning applications, web scraping, time series analysis, signal processing and more.
- Interact with Rice students from the D2K Capstone and Introduction to Data Science programs, and learn more about their projects in the poster session.
- Vote for your favorite team as they compete for an Audience Choice award!
Location: The Showcase will be held in person at Duncan Hall, Rice University. ( Map) Event Schedule (All times are in Central Time): - 5:00 pm - Introduction & Presentations | McMurtry Auditorium
- 5:30 pm - Poster Session | Martel Hall
- 6:45 pm - Awards Ceremony | McMurtry Auditorium
All students, faculty, alumni, and friends of D2K are welcome to attend! RSVP to help us order food and receive event updates for the Showcase.
Rice Launch Challenge 2022
April 20, 2022 | 5:30-9:00pm CST | The Anderson-Clarke Center’s Hudspeth AuditoriumJoin the Liu Idea Lab for Innovation and Entrepreneurship (Lilie) and support Rice University student startups at the H. Albert Napier Rice Launch Challenge (NRLC) Championships on Wednesday, April 20th! Six student finalists will pitch their ventures for the chance to win a share of $75,000+ in equity-free funding!
> REGISTER HERE
Don’t miss out on watching pitches from Rice University students, networking with entrepreneurs and innovators in Houston, and voting for your favorite venture to win the Audience Choice Award.
Attendance is open to Rice students, Rice alumni, friends of Rice, and the whole Houston entrepreneurial community!
Competition starts at 5:30PM with a reception to follow at 7PM. Seats are limited, so get your tickets while you still can!
MEET THE FINALISTS:
AutoEdge
Alfredo Costilla '22 (Innovation Fellow from Professor Xia 'Ben' Hu's DATA Lab), Daochen 'Frank' Zha '23, and Kwei-Herng 'Henry' Lai '23 are the founders of AutoEdge, an AI-powered quality assurance platform that helps small and medium manufacturers to quickly detect defects and provide clear actionable items to fix inefficiencies.
Berman Foods
Delaney Berman MBA'22 is the founder of Berman Foods, which crafts irresistible artisanal plant-based cheeses and spreads from wholesome, nutritious ingredients.
EpiFresh
Neethu Pottackal '24 (Innovation Fellow from Professor Pulickel Ajayan’s Lab) and Aasha Zinke '24 are the founders of EpiFresh, whose healthy and sustainable protein-based coating doubles the shelf-life of fruit and vegetables, reducing waste by delaying decay as it moves from the farm to your fridge.
GradGenius
David Akpakwu '23 and Chinedum Peter Ezeakacha '23 are the founders of GradGenius, designed to give the end-user a one-stop-shop experience to selecting schools based on personal interests.
Guildata
Stephanie Pons '22, Kurt Reece '22, Ryan Jensen '22 are the founders of Guildata, which provides global health organizations with data that shows the greatest return on investment, by reduction in morbidity and mortality, for public health interventions in a non-disease centric approach.
Helix Earth Technologies
Rawand Rasheed '23 (Innovation Fellow from Professor Daniel Preston’s Lab) is the founder of Helix Earth Technologies, helping save our planet by helping power plant operators reduce their plant water use and subsequently reducing their overall operating costs.
10th Eubank Conference on Real World Markets
April 25, 2022 | 8:30am CST | McMurtry Auditorium, Duncan Hall
On April 25, 10th Eubank Conference on Real World Markets will celebrate 20 years of research and education since the founding of the Center for Computational Finance and Economic Systems (CoFES).
The theme of this year’s conference is Frontiers of Quantitative Finance. Join us at this historical event, which will feature an amazing line-up of guest speakers and Rice alumni who are architects of the field, leaders in navigating the forces that shape financial markets, and innovators in using methods from the digital era to move the frontier forward.
The conference will primarily be an in-person event with some virtual lectures transmitted through a Zoom meeting. Virtual operations will be managed from Rice’s McMurtry Auditorium.
Due to the generosity of our benefactors and sponsors, there is no cost for this conference. Register here.
Texas Heart Institute Job Openings
Please see two recently posted positions in THI’s Gene Editing Lab under Dr. James Martin and Dr. Xiao Li: - Bioinformatics Analyst II
- Postdoctoral Research Associate in Bioinformatics
> Apply Here |
|
|
|
|
|
