Category Archives: News

Subash Khot: A grand vision for the impossible

Posted on by

Sources: Quanta Magazine and Simons Foundation

Subash Khot is this year’s winner of the Nevanlinna Prize. If his “Unique Games Conjecture is correct, then for many of the problems people would most like to solve, it’s hard not only to find an exact solution—finding even a good approximation of the solution is beyond the reach of a computer. This conjecture may seem, on the face of it, like a pretty parochial statement (though possibly hard to prove, or even false). Over the past decade, however, computer scientists have found to their surprise that if the conjecture is true, then a host of problems—some of them seeming to bear little resemblance to the original Unique Games problem—are also hard to approximate.” Simons Foundation. 

http://www.simonsfoundation.org/quanta/20140812-a-grand-vision-for-the-impossible/

Source: Quanta Magazine

“For the problem of coloring the nodes of a network that has a collection of constraints about which color combinations are allowed (top left), it is sometimes possible to find a coloring that satisfies all the constraints (top right). But for some networks and constraints (bottom left), it is impossible to satisfy all the constraints simultaneously. The Unique Games Conjecture concerns the problem of finding a coloring that satisfies as many constraints as possible, such as the bottom right coloring, which satisfies all the constraints except the one for the thick edge.” Quanta Magazine.

Human Brain Project under fire

Posted on by
http://www.33rdsquare.com/2013/01/human-brain-project-and-graphene-win.html

Source: 33rd square

Sources: Nature 511, July 7, 2014. Also: The Guardian, July 7, 2014.

The European Human Brain Project has come under fire. In an open letter, more than 500 neuroscientists oppose the management style of the project. The Guardian and Nature report that the international protest was triggered by the elimination of studies on cognition from the project. Cognitive neuroscientists feel ousted.

From The Guardian: “Central to the latest controversy are recent changes made by Henry Markram, head of the Human Brain Project at the Swiss Federal Institute for Technology in Lausanne. The changes sidelined cognitive scientists who study high-level brain functions, such as thought and behaviour. Without them, the brain simulation will be built from the bottom up, drawing on more fundamental science, such as studies of individual neurons.”

From Nature: “Stanislas Dehaene … says that such a simulation, “while not totally useless, will fail to elucidate brain functions and diseases, much like a simulation of every feather on a bird would fail to clarify flight”. Along with other ousted colleagues, Dehaene believes that a top-down, reverse-engineering approach is required, starting with behaviour and high-density recordings of electrical activity in the brains of humans or animals to elucidate how information is encoded and used.”

More discussion from MIT Technology Review.

3D movies of brains

Posted on by

From MIT News. With video.

Caenorhabditis_elegans

Caenorhabditis elegans. Wikimedia Commons

http://www.hhmi.org/research/zebrafish-systems-neuroscience-whole-brain-analysis-neural-circuits-underlying-learned

Larval zebrafish. Source: Howard Hughes Medical Institute

“Researchers at MIT and the University of Vienna have created an imaging system that reveals neural activity throughout the brains of living animals. This technique, the first that can generate 3-D movies of entire brains at the millisecond timescale, could help scientists discover how neuronal networks process sensory information and generate behavior. The team used the new system to simultaneously image the activity of every neuron in the worm Caenorhabditis elegans, as well as the entire brain of a zebrafish larva, offering a more complete picture of nervous system activity than has been previously possible.”

Robert Prevedel et al., Simultaneous whole-animal 3D imaging of neuronal activity using light-field microscopy. Nature Methods, 18 May 2014.

Connections: We now have the means to connect neuronal activity to behavior in a worm and in a fish. That’s exciting, but it isn’t the end of the story. Fortunately, there are a couple of passages in the recent NIH BRAIN Initiative Report that emphasize that efforts to understand the brain can’t be limited to finding the links between neuronal activity and observable behavior. “In advanced organisms our concept of ‘behavior’ must be extended to include sophisticated internal cognitive processes in addition to externally observable actions” … “Mental life can flourish within the nervous system, even if the behavioral link to the observable world is tenuous. Thus the BRAIN Initiative should focus on internal cognitive processes and mental states in addition to overt behavior.” Unfortunately, research on language and the brain doesn’t seem to be on the agenda for the NIH BRAIN initiative. That’s a big oversight, if not an outright blunder – there is probably no cognitive domain that we collectively know more about at an abstract, computational, level. There is no better window into the human mind than language, and it’s already wide open. The Max Planck Society must have seen the potential and importance of ongoing brain research on language when electing Angela Friederici as their new vice-president.  There is also an NSF report about a recent workshop on Linking Language and Cognition to Neuroscience via Computation. 

Connections: The neural code that makes us human.

Situations in Natural Language Semantics

Posted on by
http://www.gizmodo.com.au/2013/05/urville-the-metropolis-that-lives-inside-an-autistic-artists-mind/

Gilles Trehin: Urville. Source: Gizmodo

The Stanford Encyclopedia of Philosophy has recently implemented a redesign of its website. My article on Situations in Natural Language Semantics appears in a new look.

“Situation semantics was developed as an alternative to possible worlds semantics. In situation semantics, linguistic expressions are evaluated with respect to partial, rather than complete, worlds. There is no consensus about what situations are, just as there is no consensus about what possible worlds or events are. According to some, situations are structured entities consisting of relations and individuals standing in those relations. According to others, situations are particulars. In spite of unresolved foundational issues, the partiality provided by situation semantics has led to some genuinely new approaches to a variety of phenomena in natural language semantics. In the way of illustration, this article includes relatively detailed overviews of a few selected areas where situation semantics has been successful: implicit quantifier domain restrictions, donkey pronouns, and exhaustive interpretations. It moreover addresses the question of how Davidsonian event semantics can be embedded in a semantics based on situations. Other areas where a situation semantics perspective has led to progress include attitude ascriptions, questions, tense, aspect, nominalizations, implicit arguments, point of view, counterfactual conditionals, and discourse relations.”

There is a lot of recent work on domain restrictions in situation semantics, in particular on domain restrictions for definite descriptions:

Paul Elbourne’s 2002 MIT dissertation, his 2005 book on Situations and Individuals, and his 2013 book on Definite Descriptions. “My contention in this book is that definite descriptions are best analyzed semantically as expressions that contain a locally free situation variable; when the situation variable is bound or assigned a referent, the definite description ranges over or refers to individuals. So I will be working with a semantics based on situations” (from Definite Descriptions, p. 17). Elbourne also exploits situation variables for a theory of presupposition projection. 

Ezra Keshet’s 2008 MIT dissertation and his 2010 Natural Language Semantics article on Situation Economy: ” … a rule of Situation Economy is advanced, which holds that structures must have the fewest number of situation pronouns possible. Strong DPs require a situation pronoun to receive a de re reading, and therefore a restriction on the type of strong determiners is proposed, which supersedes Situation Economy in this case.”

Florian Schwarz’s 2009 UMass Amherst dissertation and his 2012 Natural Language Semantics article on Situation Pronouns in Determiner Phrases: “This paper is primarily concerned with situation pronouns inside of determiner phrases, arguing that they are introduced as arguments of (certain) determiners. Verbal predicates, on the other hand, are assumed to not combine with a situation pronoun. The various restrictions on their interpretation are shown to fall out from the semantic system that is developed based on that view.”

A blueprint for how to build a human brain

Posted on by

From a recent press release of the Allen Institute for Brain Science announcing the first major report on the  BrainSpan Atlas of the Developing Human Brain.

“Knowing where a gene is expressed in the brain can provide powerful clues about what its role is,” says Ed Lein, Investigator at the Allen Institute for Brain Science.  “This atlas gives a comprehensive view of which genes are on and off in which specific nuclei and cell types while the brain is developing during pregnancy. This means that we have a blueprint for human development: an understanding of the crucial pieces necessary for the brain to form in a normal, healthy way, and a powerful way to investigate what goes wrong in disease.”

Audio from NPR: Map of the developing brain

Visualizing genes’ role in learning and memory

Posted on by

From Kurzweil Accelerating Intelligence

“MIT bioengineers have adapted MRI to visualize gene activity inside the brains of living animals.Tracking these genes with MRI would enable scientists to learn more about how the genes control processes such as forming memories and learning new skills, says Alan Jasanoff, an MIT associate professor of biological engineering and leader of the research team.”

“The dream of molecular imaging is to provide information about the biology of intact organisms, at the molecule level,” says Jasanoff, who is also an associate member of MIT’s McGovern Institute for Brain Research. “The goal is to not have to chop up the brain, but instead to actually see things that are happening inside.”

The neural code that makes us human

Posted on by

From Science : Yosef Grodzinsky and Israel Nelken comment on Nima Mesgarani’s et al. recent finding about phonetic feature encoding in the human superior temporal gyrus.

Access from Hebrew University Website. 

“Speech representation in the auditory cortex … is governed by acoustic features, but not by just any acoustic features—the features that dominate speech representation are precisely those that are associated with abstract, linguistically defined distinctive features. Mesgarani et al., who base their investigation on linguistic distinctions, further demonstrate that features are distinguishable by the degree of the neural invariance they evoke, forming an order that is remarkably in keeping with old linguistic observations: Manner of articulation (manifesting early in developing children) produces a neural invariance that is more prominent than that related to place of articulation (manifesting late in children). A hierarchy noted in 1941 for language acquisition is now resurfacing as part of the neural sensitivity to speech sounds.”

What is so interesting about Mesgarani’s et al. finding is that they identified neural correlates of the very same phonetic features that had been posited by linguists who were stating generalizations about the sound patterns of natural languages. The pioneers in this field were Roman Jakobson and Nikolai Trubetzkoi. Jakobson and Trubetzkoi worked from their armchairs. But, with their razor-sharp analytic minds, they saw abstract patterns in natural languages. Since the patterns were so abstract, it is unlikely that they would have been discovered by neuroscientists alone. Experts on languages needed to see the patterns and develop theories of how they could be generated by a combinatorial mechanism of features. At that point, the question of neural correlates for the representation of speech sounds could be asked in a meaningful way. To be sure, Mesgarani et al. did NOT find the neural code that makes us human. That’s exaggerated. But their work is a model of how insights from linguistics might be ‘transferred’ to cognitive neuroscience.

Phonology and the brain: it’s all in the features. By Itziar Laka.

Glass Brain

Posted on by

Neuroscape Lab visualizes live brain functions using dramatic images | KurzweilAI.

GlassBrain. Credit: Neuroscape lab, UCSF

GlassBrain. Credit: Neuroscape lab, UCSF

“UC San Francisco neuroscientist Adam Gazzaley, MD, PhD, is hoping to paint a fuller picture of what is happening in the minds and bodies of those suffering from brain disease with his new lab, Neuroscape, which bridges the worlds of neuroscience and high-tech. Gazzaley aims to eliminate the need to immobilize subjects inside big, noisy machines or tether them to computers — making it impossible to simulate what it’s really like to live and interact in a complex world. Instead, in the Neuroscape lab, wireless and mobile technologies set research participants free to move around and interact inside 3D environments, while scientists make functional recordings with an array of technologies .”

Glass Brain flythrough

Representing compound words in the brain

Posted on by
http://uncmain.sites.unc.edu/files/2012/06/ccm3_030355.jpg

Source: University of North Carolina

Teon Brooks, a member of the NYU Neurolinguistics Lab, looks at the computations involved in recognizing complex words, specifically compound words (words like swan boat, bird house, book award). For semanticists like me, work on compounds by neuroscientists is particularly interesting because compounds are arguably “semantic fossils” (Jackendoff 2002). The kind of compositionality we see in compounds is more rudimentary than the full-fledged compositionally that comes with phrasal syntax, where constituents are headed by functional elements related to voice, aktionsart, aspect, tense, mood, complementizers, definiteness, quantification, and what have you. I would therefore expect the meanings of compounds like swan boat to be computed in a qualitatively different way from the meanings of phrases like boat in the shape of a swan or boat that is inhabited by swans. Is this so? Is there any evidence for this?

A village invents a language on its own

Posted on by

From Australian Geographic

Credit: Carmel O'Shannessy)

Credit: Carmel O’Shannessy

“Light Warlpiri is a newly-evolved language spoken by about 300 people in the Lajamanu Aboriginal community, located at the edge of the Tanami Desert, about 890 km south of Darwin. The language combines elements of English, Kriol (an English-based creole that emerged in the late nineteenth-century) and Warlpiri, an endangered traditional language that is spoken in central Australia by about 4000 people. The Australian linguist Dr Carmel O’Shannessy, who speaks Warlpiri, discovered the existence of Light Warlpiri while working as a teacher and linguist at Lajamanu in the late 1990s. In a paper published this month in the journal Language, Carmel documents for the first time the grammatical structure of the language and discusses the social context that led to its emergence.”

How could a group of children ‘agree’ on a language that is different from that spoken by the adults around them? In his work on conventions, the philosopher David Lewis developed an analysis of tacit agreements of this kind that was inspired by the economist Thomas Schelling.