Yonatan Dubi

Krill Prize 2017
Ben Gurion University

Dr. Yonatan Dubi

Charge and Energy Transport in Molecular devices.

A central focus of my research is the study of energy and electronic transport in molecular devices. The Field of molecular electronics aims at developing molecular-level technologies which will complement – and eventually supersede – current microelectronic technologies. For this aim, it is essential to devise molecular junctions (a single molecule or a molecular layer sandwiched between metallic electrodes) with electronic properties that mimics at least some of the behavior of modern- day components. Achieving this goal requires detailed understanding of the fundamental processes that govern the electronic transport in such molecular junctions, and using this understanding to design functional devices.

We have been studying various effects related to the transport in molecular junctions. A recent example is a model we constructed to describe a DNA-based single-molecule diode. The theory which we developed was compared to the experiments performed at the group of Prof. B. –Q. Xu  at the University of Georgia, and allowed us to explain the experimentally observed rectification (diode-like) behavior. The combined experimental-theoretical study has been published in Nature Chemistry (ref. 42). Within the same collaboration with Prof. Xu, we developed a new model for negative-differential-resistance molecular devices (refs. 34 & 38), which explained puzzling experimental results. In addition, we developed a novel theory of transport through self-assembled- monolayer-based (SAM) molecular devices (ref. 35), which was able for the first time to explain important experimental observations, such as reduction of current compared to the single-molecule junctions, molecule-length-dependence of currents, and the odd-even effect.

Another long-standing focus of interest in my group is how energy is distributed in and transferred through molecular junctions. Special focus is given to thermoelectric energy conversion – the conversion of a temperature gradient to electric voltage (e.g., refs. 12, 14-16, 18, 19, reviewed in ref. 21). Recently, we addressed the outstanding question: can a molecular system be designed which will exhibit enhanced thermoelectric performance in realistic conditions? The answer turns out to be “Yes”; we demonstrated that using novel geometries for molecular junctions (helicene- based junctions and hybrid molecule-nanoparticle junctions) one can enhance the thermoelectric performance of realistic systems (refs. 37 & 39).

Exciton transport, open quantum systems and quantum biology

Another important part of my research is the development of new methods for calculating energy transport. Borrowing ideas from the field of open quantum systems, we formulated a quantum master equation that can be used to study thermoelectricity in molecular junctions (refs. 11, 12 & 16). Recently, we extended the formulation to study, for the first time, solar energy conversion (i.e. light-to-electricity) in a quantum molecular photo-cell (ref. 36). This led us to the field which is currently a central focus of the group, namely quantum-coherent exciton transfer in photosynthetic complexes (ref. 40), where we showed that maintaining quantum coherence in model natural photosynthetic complexes leads to enhancement of power output, and provided the detailed mechanism for this enhancement. This is part of the emerging field of quantum biology, where the central question is “are there quantum coherent effects in biological systems, and if so, are they used by nature to enhance performance?”.

Yoav Goldberg

Krill Prize 2017
Bar-Ilan University

Dr. Yoav Goldberg

Description of Research

I work on developing automatic methods for understanding and generating human languages by computers (Natural Language Processing, NLP). Language is at the core of human intelligence, and is arguably what sets us apart from all other species. It is central to effective com mun icati on, as it facilitates the transfer of ideas across space and time. In the digital age, we see unprecedented growth in the scale at which individuals and institutions generate, communicate and access  information  in  written  or  spoken form. Finding ways for effectively leveraging the vast amounts of available data to discover and address people’s needs {social, professional , scientific or otherwise) is a fundamental problem in modern societies. Over the past decade we’ve seen tremendous progress in NLP capabilities, with NLP commonly facilitating breaking language barriers (machine translation), finding required information (search engines), and digesting large amounts of unstructured text into more convenient forms (summarization, information extr action). Despite this progress, the NLP problem is far from being solved, especially  when moving from types of text the systems are familiar with (mostly well edited English text ) to other genres or languages.

My research focuses on the building-blocks of NLP — the core algorithms that other language understanding systems rely on. These include computationally representing the meaning of words (lexical semantics); how words are combined to form phrases and sentences (synta x); and how this process can be reversed, inferring the structure of a sentence from its words (syntactic parsi ng).

Within lexical semantics, I worked on analyzing the foundation of “word embedding” algorithms. Word embeddings are the dominant and very popular framework for lexical semantics, and were at the time not well underst ood. In a series of works with PhD student Omer Levy we analyzed the mathematical and linguistic foundations of these  algorithms  and unified  many  algorithms  under a common  framework. This in turn allowed us to further generalize and improve the word embeddings technique. This line of work

was highly ·influentia I, accumulating over 720 citations since 2014. Products of the work are also featured in the upcoming 3rd edition of the definitive NLP textbook by Jurafsky and Martin .

Within syntactic parsing , my most recent contribution (with  PhD student  Eliyahu  Kiperwasser)  is  a parser based on bidirectional recurrent neural networks,  which  as of  April  2016 had  the  best  accuracy  in the world (the work has since been extended by  research groups  at Stanford, CMU  and Harvard)  . One of  my earlier innovations in parsing, the dynamic -oracl e training techniques, is integrated in several

widely-used parsers, including the open-source s pa Cy. i o NLP packa ge.

The field of natural-language processing interacts closely with the field  machine  learning. My  research is deeply rooted  in both fields, borrowing state-of-the-art  methods  in machine  learning and adapting them to fit the needs of linguistic tasks, leading to innovations in both machine learning and NLP. In the past two years, I became interested in the effective use of neural-n etwo rk s (a.k.a deep lea rn in g, DL) for NLP. A common trend in my work is to combine DL methods with linguistically grounded mechanisms, relying on in-depth understa nding of the DL algorithms as well as the underlying linguistic phenomena . Besides the accurate parser mentioned above, this line of work resulted in several best paper awards as detailed in my CV. I authored an extensive tutorial paper on DL techniques for NLP which is now being used in academic courses worldwide.

My focus is improving the coverage of these NLP building blocks. Current algorithms achieve impressive accuracies on well-formed English newswire text, but degrade considerably on other genres, severely limiting their applicability. I work on methods for extending their coverage to  a diverse set of textual genres (scienti fic, le gal, literary, colloquial , etc) as well as to spoken language. Accuracy also drops when attempting to work with languages with richer morphological sys t em s and freer word order than English (most of the world ‘s languages). In the Universa I Dependencies mu!ti -nati onal collaboration, we collect data on such languages under a common scheme. I recen tly started using this data to devise

DL-based lexical semantics and syntactic parsing algorithms tailored for such languages. If success ful, this will transform NLP from being applicable to a relatively narrow range of commercial scenarios to being universally applicable in a wide set of scenarios for the benefit of science and society worldwide .

Zvika Brakerski

Krill Prize 2017
Weizmann Institute

Dr. Zvika Brakerski

My Research: Advanced Cryptography via Foundational Study

Can we delegate all of our computational tasks to a remote cloud server without surrendering the privacy of the data, even to the server itself? Can we compile computer programs in a way that completely hides their inner workings, but preserves their functionality? Can we give third parties such as companies or law enforcement access to our personal data while maintaining a mathematical guarantee that they can only use it in accordance with their declared privacy policy or warrant? These questions and similar ones are becoming increasingly pertinent as our society becomes more dependent on data that is being processed by third-party algorithms. In many cases, it might appear that utility or the greater good are at odds with privacy of individuals. My current research is focused on developing new cryptographic tools for mitigating this tension, via foundational investigation of the computational properties of mathematical structures.

State-of-the-art research in cryptography, including my own prior work, suggests that even though the aforementioned wish-list appears to be paradoxical, many of these goals can be achieved using new crypto- graphic structures, in particular relying on the intractability of computational problems on high-dimensional integer lattices. In the past, I was able to make significant progress in the task of fully homomorphic encryption, which allows computation on encrypted data without decrypting it first and without leaking information, and thus addresses the first question above. However, this is merely a gateway to a new world of cryptographic applications.

The grand mission of my research endeavor is to establish cryptographic tools supporting these appli- cations and to explore the mathematical and computational framework that governs the utility/privacy interplay. My exploration is internally motivated by the following two objectives: finding the conceptually simplest structure that provides the required property, and unifying our understanding by explaining various applications as different facets of a single phenomenon. Even though practical usability and improved secu- rity are not direct objectives, I believe, and my experience shows, that orders-of-magnitude improvements in efficiency and security follow from deeper conceptual understanding. Indeed, this has been demonstrated in my work on fully homomorphic encryption, which showed how to construct this strong primitive based on a much simpler mathematical structure than previously conjectured, and in turn led to a leap in the efficiency and security of implementations.

There is still a multitude of open problems in the study of fully homomorphic encryption, as well as in the study of other advanced cryptographic primitives. These include fine-grained access control to encrypted data via attribute based encryption, and the study of program obfuscation which allows  to  “encrypt”  a program such that it is still executable, but hard to reverse engineer. The current level of understanding of these primitives varies, and the existence of some, e.g. program obfuscation, is still in the realm of conjecture (albeit with significant supporting evidence).   The study of these problems leads us time and again to    the mathematical and computational properties of high-dimension integer lattices, which can be imagined as infinite high-dimensional periodic grids. Information can be embedded in these grids in a way that is computationally hard to retrieve, and yet which enables manipulation of the information in many interesting ways. This translates into cryptographic schemes with extremely rich functionality but with a high level of security (interestingly, at the current state of knowledge, such cryptographic constructions are even immune to attacks using quantum computers, which render older cryptographic primitives insecure). The study of the cryptographic properties of lattices is fairly young, and many important pieces of the puzzle are still missing.

Our horizons are constantly expanded by the de- velopment of new techniques that we are only beginning to master, and technological developments challenge us to provide solutions to new and more intricate problems than we faced before. I feel fortunate to be a part of this great scientific effort and am very excited to see where it leads.

Asya Rolls

Krill Prize 2017
Technion

Dr. Asya Rolls

Physiological mechanisms of brain-body interactions.

The philosophical ‘mind-body problem’ examines the relationship between the mind, a mental process, and the body, a physical entity. In modern medicine, this dialog between the mental and physical state has multiple manifestations, for example, the emergence of disease following stress, or recovery in response to placebo treatment. Nevertheless, this fundamental aspect of physiology remains largely unexplored. My laboratory approaches this question by examining neuronal networks underlying specific brain-body interactions. We examine how the activity of specific brain regions, or specific neuronal ensembles correlated with defined emotional states, affects the immune system, the body’s main protective mechanism.

Using this approach, we recently uncovered a novel potential mechanism of the placebo response. For the past 50 years, most clinical trials include a placebo group to control for non-specific effects such as statistical errors, disease ontology, as well as patient expectations. This expectation of the patient for clinical improvement plays a key role in the placebo response, but it is not known how such expectation affects recovery. We used pharmacogenetics (DREADDs), a state of the art tool in neuroscience, to activate the reward system (dopaminergic neurons in the ventral tegmental area) and analyzed the effects on the immune system. We found that reward system activation strengthens the immediate, innate anti-bacterial immune response, as well as enhancing the formation of long-term immunological memory. We further showed that these effects were mediated via the sympathetic nervous system. These findings were recently published in Nature Medicine (Ben Shaanan, 2016), accompanied by a News and Views report in Nature Medicine, and were featured in Nature and Nature Neuroscience. We hypothesized that because natural rewards such as feeding and sex increase exposure to pathogens, coupling such rewarding activities with increased immunity can be of evolutionary benefit.

Our follow up study demonstrated that reward system activation can attenuate tumor growth by over 40%. We showed that these effects were mediated by an enhancement in the anti-tumor immune response, due to altered innervation to the bone marrow, the site of immune cell formation. This study proposes a potential mechanistic insight to extensive epidemiological data indicating that the patient’s mood is correlated with cancer progression (Spiegel, Lancet, 1989; Spiegel, Cancer, 2007). We expect that by uncovering the mechanisms underlying such brain-immune communication, we will be able to define means to harness the therapeutic potential embedded in the brain.

Accordingly, follow up projects in the laboratory are designed to dissect the specific effects of the peripheral nervous system on the immune compartment. We developed three strains of transgenic mice, each allowing us to control -through optogenetics- a different component of the peripheral nervous system, sympathetic, parasympathetic and serotonergic. We can now manipulate the innervation to the gut, bone marrow, lymph nodes and thymus using laser light, uncovering novel modes of communication between the nervous and immune systems.

This research direction builds on my multidisciplinary background, which places me in a unique position to investigate brain regulation of immunity. We expect this study to alter our understanding of brain-body interactions, and to set the stage for novel approaches to analyze how the state of the mind impacts physiology.

Nir Bar-Gill

Krill Prize 2017
Hebrew University

Dr. Nir Bar-Gill

Title: Quantum science and applications with diamond

 

My research aims to create a new platform for both fundamental studies in quantum science and interdisciplinary ap- plications. Specifically, I am researching the nitrogen-vacancy (NV) color center in diamond, which can serve as a building block for quantum information processing, quantum simulation and quantum sensing. My research program focuses on these three aspects of the diamond-NV platform, which I detail below.

In the context of quantum information building blocks, NV centers in diamond have been recently identified as unique atom-like quantum spins embedded in a solid-state structure, that have remarkable optical and microwave control. Over the past few years, these unique properties have propelled the field forward tremendously, resulting in significant achievements, including our demonstration of record-long coherence times approaching 1 second [7]. These results position NVs as leading candidates for various quantum information applications. We are extending the current state- of-the-art in this context through optimized, advanced control sequences [3]. Based on this, we aim at reaching the regime of fast coherent interactions in an NV ensemble (compared to the coherence time), which would allow the cre- ation of highly entangled states, such as spin-squeezed states, relevant for quantum memories and quantum metrology.

Based on these NV quantum building blocks described above, we are advancing a novel architecture for scalable quan- tum simulators in diamond. We are proceeding at the fore-front of this field, progressing toward novel hybrid ap- proaches connecting disparate physical realizations (including solid-state, photonic and atomic). In our case, we are developing quantum spin networks with a large number of individually addressed spins (optically), coupled through superconducting structures. This work could lead to scalable, universal quantum simulation and quantum computing, addressing open research questions such as frustrated spin systems, surface codes, dissipative quantum dynamics of many-body spin systems and effects of disorder on quantum and topological phases.

Finally, I am exploring the direction of using diamond devices as quantum sensors and specifically magnetic sensors, and applying this new technology to various fields of science, including physics, biology and earth sciences. For example, I am collaborating with Dr. Ron Shaar from the Institute for Earth Sciences, studying magnetic signatures of various minerals and rock samples, to gain insight into the physics of the underlying magnetization process, and to extract otherwise unavailable climactic data over a large timespan (tens of thousands of years), relevant for open ques- tions in global warming.

I am also collaborating with Prof. Stefan Hell on creating a novel two-beam super-resolution NV microscope (based on his pioneering technique, STED – STimulated Emission Depletion microscopy), which will allow sensitive pump-probe measurements on the nanoscale. This approach could reveal new information regarding magnetic and spin dynamics in condensed matter systems, e.g. measurements of the spin susceptibility of graphene and various Van-der-Walls mate- rials (including topological insulators).

Moreover, I am engaged in developing a nanoscale MRI (Magnetic Resonance Imaging) device, which could complement existing technologies, and provide a cheap, portable and yet sensitive alternative, relevant for a variety of medical sce- narios, and for situations in which traditional MRI machines are not available, e.g. third-world countries and field treat- ments.

Yossi Buganim

Krill Prize 2017
Hebrew University

Dr. Yossi Buganim

Research Vision :

Regenerative medicine is a new and expanding area that aims to replace lost or damaged cells, tissues or organs in the human body through cellular transplantation. Embryonic stem cells (ESCs) are pluripotent cells that are capable of long-term growth; self-renewal and can give rise to every cell, tissue and organ in the fetus’s body. The “hyperdynamic” chromatin state, which characterizes the ESC epigenetic state, facilitates a rapid and efficient reaction to external and internal cues that lead to the activation of key master regulators that drive the cells into their developmental fate. Thus, ESCs hold great promise for cell therapy as a source of diverse differentiated cell types. Two major bottlenecks to realizing such potential are allogenic immune rejection of ESC-derived cells by recipients and ethical issues. In 2006, two Japanese scientists, Takahashi and Yamanaka, changed the way we used to think about cell plasticity when they showed that introduction of four transcription factors, Oct4, Sox2, Klf4 and Myc (OSKM), can reprogram adult fibroblasts into functional embryonic stem-like cells (also termed induced pluripotent stem cells (iPSCs)). The notion that as little as four factors are sufficient to reset the epigenome of a cell, opened a new avenue where scientists have attempted to convert different adult cells into other somatic cell types from ontogenetically different lineages, by avoiding the pluripotent state, using a specific subset of key master regulators. Several subsets of cell types such as hematopoietic cells, different neuronal cells cardiomyocytes, hepatocytes, embryonic Sertoli cells, endothelial cells and RPE were converted from different somatic cells by employing the direct conversion approach. Importantly, the generation of iPSCs and directly converted cells resolve the problems of allogenic immune rejection of ESC-derived cells by recipients and ethical issues. However, the majority of the directly converted cells are not stable and represent mostly incomplete reprogramming state and the vast majority of the resulting iPSCs, although could activate their endogenous circuitry, exhibit poor developmental potential in mice. This suggests that the current prevailing reprogramming methods are not ideal and must be improved before considering applying the converted cells and iPSCs in the clinic.

 

To overcome these barriers, my lab is focused on identifying and investigating the components (e.g. transcription factors, chromatin remodelers, epigenetic marks and metabolites) that determine cellular identity and the elements that can force cells to acquire an alternative fate. To that end, we study two major cell fate decision processes in the embryo: (i) Embryo vs placenta [i.e. inner cell mass (ICM) vs trophectoderm (TE)] and (ii) sex determination during gonadal development [i.e. Sertoli cells (male) vs Granulosa cells (female)]. To identify key elements that play a role in these two fascinating processes we take advantage of our expertise in nuclear reprogramming, embryo manipulation and single-cell technologies, to generate engineered cells that can be captured and examined during embryonic development and following various nuclear reprogramming processes driven by defined factors. We study four somatic cell conversion models, which recapitulate, at least partially, the changes that occur in vivo during the above mentioned two cell fate decision processes: (i) fibroblasts into induced pluripotent stem cells [(iPSCs), an in vitro equivalent to ICM], (ii) fibroblasts into induced trophoblast stem cells (iTSCs), an in vitro equivalent to TE], (iii) induced embryonic Sertoli-like cells [(ieSCs), an in vitro equivalent to embryonic Sertoli cells] and (iv) induced embryonic Granulose cells (ieGCs), an in vitro equivalent to embryonic Granulosa cells]. Our overarching goal is to uncover and explore the elements that are shared among the in vivo and in vitro processes, to shed light on basic cell fate decision processes during embryonic development and to decipher the mechanisms underlying intact somatic nuclear reprogramming. Highlighted elements (e.g. gene, enhancer, miRNA, LincRNA and metabolite) that are discovered by the single-cell analyses are eventually subjected to gain and loss of function experiments within a developing embryo, using gene editing techniques such CRISPR/Cas9, to validate the in vitro results. Pursuing these goals will enable the generation of high quality converted cells and iPSCs for regenerative medicine and fertility treatment.

Keren Censor-Hillel

Krill Prize 2016
Technion

Dr. Keren Censor-Hillel

Research Topic: The computational power of distributed computing models

As a constantly increasing number of technologies involve computing on a large scale, manyof them rely on distributed systems. These systems consist of multiple computing components that do not rely on centralized control. Examples of paramount importance
include huge networks or databases used by large companies, as well as social or economic networks. The significance of such systems makes it vital that we understand the fundamental theory that underlies distributed computation. The goal of my research is to unify and compare models of distributed computation. In other words, my high-level goal is to provide exact measures that separate hardness that is attributed to the system
weaknesses from the difficulties that arise from each specific problem.

The main paradigm for distributed systems is considering them as graphs. The computing
components are represented by vertices and the communication links by edges. The
complexity measure of a distributed algorithm that we aim to minimize is the amount of
communication required in order to execute a given task.

A huge amount of research has been devoted to understanding computation in the above
basic setting (the Local model). This model is aimed at abstracting away many uncertainties that arise in distributed computing, such as lack of global clocks, faulty components, or errors in communication. Because this model is clean and captures some of the main challenges that distributed computing faces, significant progress has been made and the complexity of many computations is fairly well-understood.

In my research, I address the fundamental question of what happens to the computation
when practical limitations are brought back into the picture.One example for this is communication restrictions, such as a limit on the size of messages (the Congest model) or on the congestion at each node (the Broadcast-Congest model). My previous results provide algorithms for globally disseminating information in the Congest and Broadcast-Congest models as well as lower bounds for doing so. The methods used include novel distributed connectivity decomposition algorithms and analysis of connectivity after
randomly sampling a graph. I also study the robustness of algorithms in these models against failures, with some new randomized techniques presented by a graduate student of mine and myself.

A second line of study I pursue is the Clique model of computation, in which the
communication graph is complete, and is separated from the graph whose attributes are
required to be studied. This model attempts to separate the hardness of computation that
arises from distances in the communication graph, from that which is caused by congestion over the links. Along with a graduate student of mine and collaborators, I recently obtained an exciting result that gives us ability to use algebraic techniques for distributed computations in this model. Our ultimate goal is to use these techniques for solving problems in the more realistic but harder Congest model, by simulating the Clique in dense parts of the network graph. A third example for practical limitations on distributed computing is the possibility of errors in communication. This can be caused by collisions in a wireless network, or by any other type of noise or fault in the system. An important tool that comes into play is the ability to use coding for the information that is sent. This constitutes another line of research that I pursue, in which I provided coding techniques for additive wireless networks, and I plan to study coding for additional settings that are prone to errors. My work provides an important complement to the combined efforts of the distributed computing community to study and understand particular distributed tasks. Moreover, my work typically has impact on theoretical computer science in general. One example is my work in the Broadcast-Congest model of computation, which led us to obtain bounds on the vertex connectivity of subgraphs induced by sampling, a question of high importance to graph theory.

Netanel Lindner

Krill Prize 2016
Technion

Netanel Lindner 

Research topic: expanding the topological frontier in quantum matter.

Topological phases exhibit a fascinating interplay of quantum mechanics and physics of many particles, and are possibly the most extreme form of quantum phases of matter. They bring with them not only a new paradigm, deviating from the conventional one that classifies phases of matter through symmetry breaking, but also new types of particle statistics, and the possibility to encode quantum information in a form which is resilient to interaction with the environment. Thanks to these fascinating properties, they stand at the forefront of fundamental research in condensed matter physics, and concurrently considered to be attractive candidates for future platforms for performing quantum computation tasks. This scientific beauty, however, is not uncovered easily: topological phases of matter are challenging to work with both on the theoretical and experimental front. Much has yet to be learned and understood about the theory underlying these phases of matter; the challenge of experimentally harnessing them is yet at an even more infantile stage.

Topological phenomena has been playing an increasingly important role in our understanding of the collective quantum behavior of electrons. How wide is the role of topology in condensed matter physics? Can we find new regimes in which topological phenomena are exhibited? Can we engineer experimentally accessible condensed matter systems exhibiting this myriad of phenomena, and how can we probe them? These are some of the questions that are dealt within the scope of my research program. My aim is to find new manifestations of topological phenomena in quantum matter; propose routes for realizing them; and develop and optimize methods to probe, control, and utilize them.

Netanel Lindner focuses his attention on two main directions: non-Abelian quantum phases of matter and topological phenomena in non-equilibrium systems.