- Research News
- Contact Us
- Department Life
- Oral History of Cornell CS
- Department Timeline
- Annual Reports
- Job Postings
- Ithaca Info
- Internal info
- CS Colloquium
- CU Tech Colloquium
- BOOM 2017
- SoNIC Workshop
- Conway-Walker Lecture Series
- Salton Lecture Series
- Seminars / Lectures
- Big Red Hacks
- 50th Anniversary Symposium
- Cornell University High School Programming Contest
- Game Design Initiative
- SJTU International Workshop
- Seminars & Lectures
Quantum Computing and the Limits of the Efficiently Computable
I'll discuss the capabilities and limits of quantum computers, the evidence for phenomena in Nature that are intractable
Computer Graphics Research at Pixar
Pixar is widely known for its animated films. What's less widely known is the foundational research that Pixar did to made those films possible. That research was greatly supported by two types of connections. One was having close ties to university research - the early researchers at Pixar came out of universities and continued to publish and be part of the academic community. The other was working with artists - meeting the practical needs of our artists often ended up requiring fundamental advances in theoretical understanding. Furthermore, the advances made for entertainment frequently turned out to be crucial for other applications as well, such as medical imaging and GPUs. This talk gives an overview of research at Pixar and the importance of our ties to universities and our collaboration with artists.
A Surprising Application of Differential Privacy
A great deal of effort has been made to reduce the risk of spurious scientific discoveries, from the use of holdout sets and sophisticated cross-validation techniques, to procedures for controlling the false discovery rate in multiple hypothesis testing. However, there is a fundamental disconnect between the theoretical results and the practice of science: the theory assumes a fixed collection of hypotheses to be tested, or learning algorithms to be applied, selected non-adaptively before the data are gathered, whereas science is by definition an adaptive process, in which data are shared and re-used, and hypotheses and new studies are generated on the basis of data exploration and previous outcomes.
Surprisingly, the challenges of adaptivity can be addressed using insights from differential privacy, a field of study supporting a definition of privacy tailored to private data analysis. As a corollary we show how to safely reuse a holdout set a great many times without undermining its power of ``correctness protection,'' even when hypotheses and computations are chosen adaptively. Armed with this technique, the analyst is free to explore the data ad libitum, generating and evaluating hypotheses, verifying results on the holdout, and backtracking as needed.
A Research Institute as Start-up
The Institute of Science and Technology Austria (IST Austria) recently celebrated its fifth birthday. In its first five years, the Institute has grown from 0 to 30 research groups in biology, physics, mathematics, and computer science. During the next decade, the Institute is expected to triple in size, all the time focusing on its twin mission of performing basic research in science and educating doctoral students. It does not happen often that a new scientific institution is founded. We present the decisions that have gone into the development of IST Austria, which we believe will pave the way for the Institute to join the league of leading research institutions in the world.
Some Challenges in Industrializing IT Services
World-wide Enterprise IT spend on products and services is expected to touch $3.8 trillion dollars this year and it continues to grow faster than global GDP. Over the past three decades the Indian IT services industry has played a critical role in using technology to create business differentiation for most global 2000 companies. The industry increasingly resembles the manufacturing industry at the turn of the 20th century - with challenges in operating at a large scale, defining discrete tasks, and managing mature processes. However, there is as yet insufficient focus on productivity, a key ingredient in industrializing any service. In this talk I will highlight some of the key challenges that arise when industrializing IT services and in particular I will focus on the problem of measuring individual productivity. I will argue why this is a hard problem; why it requires a confluence of effort from a variety of research areas within computer science; and why addressing this problem has the potential to re-organize how technology is delivered and managed in the enterprise of the future.
One robot for every task
Advances in computing are transforming the way we live, work, communicate, and learn, have driven advances in nearly all other fields and have powered the economy. The digitization of practically everything coupled with advances in the mobile Internet, the automation of knowledge work, and advanced robotics promises a future with democratized use of machines and wide-spread customization.
25 years ago John Hopcroft asked whether we could have a robot pour coffee in a coffe cup. Today we can finally answer yes. In robotics, increasingly more capable robots with enhanced dexterity, mobility, and sensing have the potential of wide-spread impact, from manufacturing and construction to health-care to smart cities and transportation and everyday life. In this talk we will discuss recent technologies to rapidly design, customize, fabricate, and use robots.