Past progress report: December 2011.

Here’s what the Singularity Institute did in January 2012:

• Winter fundraiser: We continued raising funds in January, but we still have about $30,000 left to go in our winter fundraiser before the deadline of February 20th. Please support our recent efforts toward greater transparency, efficiency, and productivity by donating now!
• Strategic discussions: In January we held a long and ongoing series of discussions concerning Singularity Institute strategy. Which scenarios are the most probable “desirable” futures for humanity, which ones can our species influence most significantly, and which ones should the Singularity Institute work to influence? Which tactical moves should the Singularity Institute make right now? How can our efforts best create synergies with other organizations focused on existential risks? These are complex questions, and in January, Singularity Institute staff members spent dozens of hours sharing their own evidence and arguments. (At one point, we also called upon the expertise of more than a dozen elite mathematicians in our circle.) These discussions continue today, and our opinions on strategy appear to be more unified than they were at the beginning of the month. But there is more evidence to gather and more strategic analysis to be done.
• Ongoing long-term projects: Amy continued her preparations for Singularity Summit 2012. Michael Anissimov and others continued work on the Singularity Institute’s new website, which will feature loads of new content and a cleaner design. As part of our transparency efforts, Luke gave a second Q&A about the Singularity Institute, an interview at 80,000 Hours, and another interview at Singularity 1 on 1. Louie continued to work on improving our book-keeping and accounting practices. Anissimov finished thanking all donors who gave during 2011. (If you donated in 2011 and were not thanked, please contact michael@singularity.org!)
• Articles: Luke and Anna continued writing “Intelligence Explosion: Evidence and Import,” and Carl continued working with Stuart Armstrong of FHI on “Arms Races and Intelligence Explosions.” Luke began adding non-English translations at Facing the Singularity, and published No God to Save Us and Value is Complex and Fragile there. Carl, with co-author Nick Bostrom, submitted a final version of “How Hard is Artificial Intelligence?” to the Journal of Consciousness Studies. For Less Wrong, Luke published What Curiosity Looks Like, Can the Chain Still Hold You?, Leveling Up in Rationality, and The Human’s Hidden Utility Function (Maybe); Anna published Urges vs. Goals. Eliezer continued work on his new Bayes Tutorial. Luke and Anna wrote a report on the workshops that followed Singularity Summit 2011, which should be published soon.
• Rationality Group: Anna continued to lead the development of a new rationality education organization, temporarily called “Rationality Group.” Per our strategic plan, we will launch this new organization soon, so that the Singularity Institute can focus on its efforts on activities related to AI risk. In January we made one trial-hire for the new organization, and reached out to dozens of other potential team members. We also published a draft of one rationality lesson as a sample (PowerPoint slides + booklet PDFs).
• New team members: Kevin Fischer of GK International joined our board of directors. We also added several new research associates: Paul Christiano, Tyrrell McAllister, János Kramar, and Mihaly Barasz (at Google Switzerland). Luke hired an executive assistant, Denise Simard.
• Meetings with advisors, supporters, and potential researchers: As usual, various SI staff met or spoke with dozens of advisors, supporters, and collaborators about how to build the existential risk community, how to mitigate AI risk, how to improve the Singularity Institute’s effectiveness, and other topics. We also met with several potential researchers to gauge their interest and abilities.
• Unpublished research: As with most research institutes, most of our research is unpublished while still informing our views and arguments. Unpublished research in January 2011 included research on CEV, decision theories, rationality training, population ethics, embryo selection, the neurobiology of preference, the psychology of goals, and more.
• Relaunched the Visiting Fellows program: In January we relaunched our Visiting Fellows program. Instead of hosting many visiting fellows at once, we will now host only 1-2 fellows at a time, for a limited duration unique to each visiting fellow. Our visiting fellow for the last week of January was Princeton philosophy undergraduate Jake Nebel. If you’re interested, please apply to our Visiting Fellows program here.
• Much more: Launched a redesign of HPMoR.com, continued work in the optimal philanthropy movement, continued work on our first annual report, and much more.

Finally, we’d like to recognize our most active volunteers in January 2012: Mitchell Owen, Huon Wilson, David Althaus, Florent Berthet, Sergio Terrero, “Lightwave,” Emile Kroeger, and Giles Edkins. Thanks everyone! (And, our apologies if we forgot to name you!)

Understanding the human mind is the key to social robotics, and researchers describe what we can expect from this field in the future.

We present an extensive evaluation of 17 confidence measures for stereo matching that compares the most widely used measures as well as several novel techniques proposed here. We begin by categorizing these methods according to which aspects of stereo cost estimation they take into account and, then, assess their strengths and weaknesses. The evaluation is conducted using a winner-take-all framework on binocular and multi-baseline datasets with ground truth. It measures the capability of each confidence method to rank depth estimates according to their likelihood for being correct, to detect occluded pixels and to generate low-error depth maps by selecting among multiple hypotheses for each pixel. Our work was motivated by the observation that such an evaluation is missing from the rapidly maturing stereo literature and that our findings would be helpful to researchers binocular and multi-view stereo.

The path following algorithm was proposed recently to solve the matching problems on undirected graph models, and exhibited a state-of-art performance on matching accuracy. In this paper we extend the path following algorithm to the matching problems on directed graph models, by proposing a concave relaxation for the problem. Based on the concave and convex relaxations, a series of objective functions are constructed, and the Frank-Wolfe algorithm is then utilized to minimize them. Several experiments on synthetic and real data witness the validity of the extended path following algorithm.

In this paper we present an efficient new approach for addressing two-view minimal-case problems in camera motion estimation, most notably the so-called five-point relative orientation, and the six-point focal-length problem. Our approach is based on the hidden variable technique for solving multivariate polynomial systems. The resulting algorithm is conceptually simple, which involves a relaxation which replaces monomials in all but one of the variables to reduce the problem to the solution of sets of linear equations, and finding the solution of a polynomial eigenvalue problem. To actually solve the polynomial eigenvalues efficiently, we make novel use of several numeric techniques, which include quotient-free Gaussian elimination, Levinson-Durbin iteration, as well as a dedicated root-polishing procedure. We have tested the approach on different minimal cases and extensions, with very satisfactory results obtained. Both executables and source codes of the proposed algorithms are made online and freely downloadable.