Monthly Archives: October 2015

Shakhnarovich in CS Thurs. 10/29

Greg Shakhnarovich of the University of Chicago will present “Zoom-out Features For Image Understanding” in the Machine Learning and Friends lunch Thursday Oct. 29 at 1 pm (arrive at 12:45 for pizza). An abstract follows.

I will describe a novel feed-forward architecture, which maps small image elements (pixels or superpixels) to rich feature representations extracted from a sequence of nested regions of increasing extent. These regions are obtained by “zooming out” from the superpixel all the way to scene-level resolution. Applied to semantic segmentation, our approach exploits statistical structure in the image and in the label space without setting up explicit structured prediction mechanisms, and thus avoids complex and expensive inference. Instead superpixels are classified by a feedforward multilayer network with skip-layer connections spanning the zoomout levels. Using off-the-shelf network, pre-trained on ImageNet classification task, this zoom-out architecture achieves near state-of-the-art accuracy on the PASCAL VOC 2012 test set.

Agreement Workshop starts Thursday 10/29 at 9 am

The Agreement Workshop begins this Thursday at 9 am in ILC with a talk from Brian Dillon of UMass Linguistics. The full program, which lasts all day Thursday and Friday and features a wide range of experimental and theoretical research on syntactic agreement by local and visiting speakers, is here. Feel free to come for any of it, even if you haven’t registered. (Rumor has it that the participants will be joining local linguists to eat local apples Friday night; please ask Brian or Joe Pater for details if this is news to you and you’d like to join us too).

Halpert in Linguistics Fri. 10/23 at 3:30

Claire Halpert of the University of Minnesota will give a talk on Friday, Oct. 23, in ILC N400 at 3:30. (The title and abstract follow.)

Escape clause

In this talk, I investigate the syntactic properties of clausal arguments, looking in particular at whether A-movement is permitted out of finite clauses and at whether these clauses themselves may undergo movement or establish agreement relationships. In English, argument clauses show some puzzling distributional properties compared to their nominal counterparts. In particular, they appear to satisfy selectional requirements of verbs, but can also combine directly with non-nominal-taking nouns and adjectives. Stowell (1981) and many others have treated these differences as arising from how syntactic case interacts with nominals and clauses. In a recent approach, Moulton (2015) argues that the distributional properties of propositional argument clauses are due to their semantic type: these clauses are type e,st and so must combine via predicate modification, unlike nominals. In contrast to English, I show that in the Bantu language Zulu, certain non-nominalized finite CPs exhibit identical selectional properties to nominals, therefore requiring a different treatment from those proposed in the previous literature. These clauses, also like nominals, appear to control phi-agreement and trigger intervention effects in predictable ways. At the same time, these clauses differ from nominals (and nominalized clauses) in the language in certain respects. I will argue that these properties shed light on the role that phi-agreement plays in the transparency/opacity of finite clauses for A-movement and on the nature of barrier effects in the syntax more generally.

Potter in Cognitive Brown Bag Weds. 10/21 at noon

Mary C. Potter, Professor of Psychology Emerita in the MIT Department of Brain and Cognitive Sciences will present “Detecting picture meaning in extreme conditions” in the Cognitive Brown Bag Wednesday at noon in Tobin 521B. An abstract follows.

Abstract. Potter, Wyble, Hagmann, & McCourt (2014) reported that a new pictured scene in an RSVP sequence can be understood (matched to a name) with durations as brief as 13 ms/picture. Although  d’ increased as duration increased from 13 ms to 27, 53, and 80 ms/picture and was higher when the name was given before than after the sequence, it was above chance at all durations, whether the name came before or after the sequence. I will describe this and subsequent research that replicated and extended those results, including recent studies using spoken vs. written names, with very tight timing between the onset of the name and the onset of the RSVP pictures. Whether these results indicate feedforward processing (as we suggest) or are accounted for in some other way, they represent a challenge to models of visual attention and perception.

Freeman CS Distinguished Lecture, Weds. 10/21 at 4 pm

Bill Freeman of MIT will present “A Big World of Tiny Motions” in the Distinguished Lecture series, Wednesday, October 21, 2015 Computer Science Building, Room 151 from 4:00pm to 5:00pm. A reception will be held in the Atrium at 3:40 pm.

Abstract. We have developed a “motion microscope” to visualize small motions by synthesizing a video with the desired motions amplified. The project began as an algorithm to amplify small color changes in videos, allowing color changes from blood flow to be visualized. Modifications to this algorithm allow small motions to be amplified in a video. I’ll describe the algorithms, and show color-magnified videos of adults and babies, and motion-magnified videos of throats, pipes, cars, smoke, and pregnant bellies. The motion microscope lets us see the world of tiny motions, and it may be useful in areas of science and engineering.

Having this tool led us to explore other vision problems involving tiny motions. I’ll describe recent work in analyzing fluid flow and depth by exploiting small motions in video or stereo video sequences caused by refraction of turbulent air flow (joint work with the authors below and Tianfan Xue, Anat Levin, and Hossein Mobahi). We have also developed a “visual microphone” to record sounds by watching objects, like a bag of chips, vibrate (joint with the authors below and Abe Davis and Gautam Mysore).

Collaborators: Michael Rubinstein, Neal Wadhwa, and co-PI Fredo Durand.

For project web pages and radio segments, visit the events page.

TED or TEDx talks by students:

See invisible motion, hear silent sounds
How a silent video can reveal sound: Abe Davis’ knockout tech demo at TED2015

Bio: William T. Freeman is the Thomas and Gerd Perkins Professor of Electrical Engineering and Computer Science at MIT, and a member of the Computer Science and Artificial Intelligence Laboratory (CSAIL) there. He is currently on a partial leave from MIT, starting a computer vision group at Google in Cambridge, MA.

His current research interests include machine learning applied to computer vision, Bayesian models of visual perception, and computational photography. He received outstanding paper awards at computer vision or machine learning conferences in 1997, 2006, 2009 and 2012, and test-of-time awards for papers from 1990 and 1995. Previous research topics include steerable filters and pyramids, orientation histograms, the generic viewpoint assumption, color constancy, computer vision for computer games, and belief propagation in networks with loops.

He is active in the program or organizing committees of computer vision, graphics, and machine learning conferences. He was the program co-chair for ICCV 2005, and for CVPR 2013.

Cognitive Science Grant Writing Group?

Could the Cognitive Science Initiative help you to get (more) external funding? Would bi-monthly meetings with your colleagues over lunch help you to meet grant deadlines? Would you like assistance finding successful grant writers to provide comments on your proposals? Do you need help identifying the collaborators who would make your proposal truly interdisciplinary? Would you like support to invite a potential mentor from another institution? If your answer to any or all of those questions is Yes, or Maybe, or Perhaps if … please send a quick email to Lisa Sanders (lsanders@psych.umass.edu) indicating your interest and the type of grant writing help that you would find most useful. With sufficient interest, we’ll make it happen.

Mahadevan Spring 2016 Course: Building a Deep Mind in the 21st century

Sridhar Mahadevan of Computer Science will be offering a graduate seminar in Artificial Intelligence this spring.

SPRING 2016: COMPSCI 791DM.: Building a Deep Mind in the 21st century

Many cognitive abilities once solely the province of biological systems are now routinely achievable with machines. Once no more than a dream, abilities such as complex 3D perception, natural language, machine translation, speech recognition, learning, and reasoning are now routinely achievable even on low cost hardware, such as cellphones, tablets,  or similar devices, which can access larger scale servers for offline computation.  This seminar explores the next frontiers for AI, presuming that many human cognitive abilities will be largely achieved in the next decade (vision, language, learning, reasoning, speech recognition), with the massive computing and data resources likely to be available to individuals and to corporations like Google, Baidu, Facebook, and IBM. To put it another way, core cognitive abilities are no longer “mysteries”. but matters of mundane albeit challenging engineering.

What’s left for AI? Will “artificial intelligence” be achieved when machines achieve human level performance in these core cognitive abilities, or, as one philosopher put it in the early heady days of AI, have we merely climbed a tall tree on the way to getting to the moon? Our thesis is that the really interesting aspects of AI are only now beginning to be possible, given that the “operating system” level functionalities listed above are achievable. This seminar will discuss how AI research in the next decade or two can begin to tackle each of these problems.

1. Emotion: current AI systems can learn effectively from reinforcement, but feel no emotion. Is emotion necessary to build truly intelligent cyborgs? We explore the current theories of emotion, and its connections to rational decision making.

2. Consciousness: AI systems can carry out inference and reason about choices, but are not conscious of their own self, in the sense that most mammals are. What’s missing? We review the current theories of consciousness.

3. Curiosity: AI systems can be programmed to carry out tasks using high level rewards and goals, but seem incapable of setting themselves their own tasks and goals. How can we build AI systems that are curious in the sense that children are. What are the essential elements of curiosity?

4. Mortality: AI systems have no fear of “death” and do not concern themselves with their “mortality”, in the sense that humans (and presumably other mammals) do. However, as AI systems learn from experience and compile massive knowledge bases, they may face similar challenges in being “unplugged”, faced with the loss of everything they know and have learned. (HAL in 2001: A Space Odyssey certainly feared its own death). Why is mortality important to incorporate in future AI systems?