Yearly Archives: 2019

New Fund Helps US-Israel Research Ties Drive the Global Artificial Intelligence Revolution

The Zuckerman Fund for Interdisciplinary Research in Machine Learning and Artificial Intelligence taps into talent at multiple campuses of Israel’s AI research leader, Technion – Israel Institute of Technology

With the fast-growing Artificial Intelligence (AI) sector and its associated fields poised to change the world as we know it, a newly launched fund is leveraging the power of U.S.-Israel collaboration to produce groundbreaking interdisciplinary research that seeks to place Israel at the center of the global AI map.

Introduced today during an event at the Technion – Israel Institute of Technology in Haifa, the Zuckerman Fund for Interdisciplinary Research in Machine Learning and Artificial Intelligence will support up to six research projects annually that are conducted between researchers affiliated with at least two faculties at the Technion’s main campus in Israel as well as the Jacobs Technion–Cornell Institute in New York. The projects will focus on applied research in the sectors which are increasingly influenced by AI, including health and medicine, financial technology, autonomous vehicles, home, and industrial robots, smart environments, agricultural technology, and defense and homeland security.

The Technion is home to the greatest proportion (28 percent) of the 270 researchers in the fields of AI, data science, and smart robotics working across Israel’s universities and colleges, according to a recent report compiled by the Samuel Neaman Institute for National Policy Research at the request of the Israeli Ministry of Science’s National Council for Research and Development. The report found that the top three interests of the Israeli researchers are machine learning, AI, autonomous systems, and smart robotics, with their other interests spanning human-computer interaction, natural language processing, computer vision, data mining, distributed systems/computing, big data, multi-agent systems, neural networks, and reinforcement learning. 

“We are utilizing both the domestic and global resources of Israel’s AI research leader, the Technion, as a game-changing platform for positioning the ‘start-up nation’ at the forefront of the global AI revolution,” said James Gertler, a trustee of the Zuckerman Institute and Zuckerman Family Foundation, a founding supporter of the AI fund. “The time for this fund is now because the future is now.”

 “Supporting tomorrow’s technological solutions through today’s research is a crucial manifestation of our vision that philanthropy’s ultimate goal is the betterment of society as a whole,” said Eric Gertler, a trustee of the Zuckerman Institute. “Joint innovation in AI and other high-tech areas also forms the bedrock of the next frontier in the U.S.-Israel relationship. We hope that others follow our lead in paving the way for stronger American-Israeli ties through collaborative research.”

Prof. Boaz Golany, Technion’s Vice President for Resource Development and External Relations who also coordinates the Zuckerman STEM Leadership Program on behalf of the Technion, Tel Aviv University, The Hebrew University and the Weizmann Institute added: “we are grateful to our partners at the Zuckerman Institute for their continuous support of our pursuit of academic excellence.  By focusing on AI, an area in which the Technion already has a global leadership position, the Zuckerman Institute will play an important role in boosting our research capabilities in AI.” “I am certain that the fruits of the research projects that will be supported through this new fund will yield great benefits not only to Israel and its major ally – the USA – but also to the rest of the world” stated Prof. Golany.

Mr. Oren Gershon, Inference Group & Site Manager, Artificial Intelligence Products Group (AIPG) at Intel, thanked the Zuckerman Institute for their support of the ML&IS Center where Intel has been a partner over the last year and added that given the significant research results already achieved in the last few months, Intel has decided to extend its support to the program (originally committed to 18 months) for another period well into 2020 .        

About the Zuckerman Fund for Interdisciplinary Research in Machine Learning and Artificial Intelligence

The Zuckerman Fund for Interdisciplinary Research in Machine Learning and Artificial Intelligence is an initiative of the Zuckerman Institute, which was created for charitable, scientific, literary and educational purposes. Mort Zuckerman is an American businessman, who serves as Chairman Emeritus of Boston Properties — one of the largest real estate companies in the United States. Mort Zuckerman launched the Zuckerman STEM Leadership program in 2016 to support future generations of STEM leaders in Israel and the United States and build strong academic bonds between the countries.

Welcome to the 2019 President’s Report, in which we review a decade of progress and look forward to continuing fruition of the Technion vision.

When I received the tremendous honor of becoming Technion President ten years ago, I felt that I was handed an enormous responsibility: not only to maintain Technion’s status as Israel’s first and leading university but to steer it onward and upward to new levels of excellence and global standing.

Fast forward with AI, EE, Bioprinting and so much more at Technion Israel. Enjoy and subscribe to Technion LIVE!

“Supporting tomorrow’s technological solutions through today’s research is a crucial manifestation of our vision that philanthropy’s ultimate goal is the betterment of society as a whole,” added fellow trustee Eric J. Gertler.

Design Meets Technology in the First Design-Tech International Conference at the Technion, Israel

“Design-Tech 2019”, the first International Design and Technology Conference hosted at Technion Israel, took place on 18-19 June 2019. The conference attracted over 150 participants and showcased over 40 featured presentations of innovative projects by designers and researchers.

Ezra Ozery who got the audience on their feet with their hands in the air during his presentation on “The Gamification of Physical Therapy in Virtual Reality”

The 2-day gathering was an inspiring, diverse and interesting platform for sharing ideas, knowledge, and questions to improve and deepen the understanding of the links between design and technology in the present era.

The conference was initiated by Prof. Ezri Tarazi, Head of the Industrial Design Graduate Program at the Faculty of Architecture and Town Planning In his opening words, Prof. Tarazi spoke of the meeting point of design and technology: “We urgently need to move fast towards humane-centered design and nature-centered design. We need to shift quickly our minds, awareness, values, and goals,” he said.

Keynote Speakers were Prof. Dr. Gesche Joost, from the Design Research Lab at the Berlin University of the Arts – who spoke of “Civic Design for an Inclusive Digital Society”; and Prof. Kalevi “Eetu” Ekman from Aalto Design Factory, Finland, who shared his thoughts on “Passion based Learning”.

Prof. Dr. Gesche Joost

Prof. Barry Katz, from the California College of the Arts and the Stanford University Design Group said that: “the last 10 years have seen deeper and more far-ranging changes than any previous decade in human history”. The keynote speaker then posed the question: Do the theories and methods that designers have applied so successfully in the past provide give a reliable guide to the future?”

The conference attracted leading Israeli designers and researchers, such as Tamar Yehezkel, Yaniv Glozman, Michal Greenberg Abrahami, Tami Warshavski, Safi Hefetz, Merav Perez, Arielle Blonder, Shira Shoval, Romi Mikulinsky, Matan Zohar, Alon Shikar, Raz Elias, Liron Lavi Turkenich, Yonatan Assouline and Asaf Cohen.

Technion students and researchers presenting their work included: Ezra Ozery who got the audience on their feet with their hands in the air during his presentation on “The Gamification of Physical Therapy in Virtual Reality”; Alexander Geht, who presented his initiative “Anatomic Mass Customization and Assistive Technology”; Avishag Shemesh who shared her investigation of the question “How does Geometry of Space Affect our Emotions”; Ofer Berman spoke about 3D printing of “Digital Tabular Coral”; Offri Lotan presented “Re-Design the Modern Surfboard”; Dr. David Behar shared research with Ronen Eidelman on “Community Surveillance and Mass Surveillance”; and Tom Shaked presented his work on “Autonomous Robotic Stone Dressing”.

From overseas, the conference hosted Alessia Buffagni and Francesca Toso from Italy; Deepshikha from India; Denisa Reshef Kera from Spain; Philip Meier, Ute Hilgers-Yilmaz, Bianca Herlo, from Germany; and Isabella Molloy from New Zealand. Israeli Marine Zorea presented her Japanese research on “Sonic Mapping of Everyday Interactions at the Elderly Home”; and Indian Kartikeya Date presented his Israeli research on “Designing Archives in The Digital Age”.

Prof. Wendy Ju – a special guest from the Jacobs Technion Cornell Institute in New York City – gave a presentation on: “Addressing Cross-Cultural Differences in Design”.

Design Tech 2019  was managed by Valeria Geselev, produced by Meital Gotfrid and Limor Abas Or, and supported by The Firefly Scientists’ Foundation.

All the conference lectures were documented, and are to be shared on the Design-Tech YouTube Channel and Website designtech.net.technion.ac.il

Photography by Lauren Blumenthal.

The first detection of an interstellar asteroid/comet-like object visiting the Solar system two years ago has sparked the ideas about the possibility of interstellar travel. New research from the Technion–Israel Institute of Technology suggests that such objects also raise far-reaching implications about the origins of planets across the galaxy, and possibly even the initial formation of the Solar system itself. 

The asteroid/comet-like object named “’Oumuamua” confirmed decades-old scientific expectations that suggested that the interstellar medium is full of loose chunks of rock flying around. Such debris was thought to be ejected from planetary systems in the aftermath of planet formation, when large planets formed and kicked out some of the leftover minor-planets and planetesimals still lying around.  From time to time, some fraction of these ejected rocks can still encounter foreign stars. In fortunate instances, this phenomenon can be observed as it swooshes through the Solar system.

The Technion researchers, Evgeni Grishin, Hagai Perets and Yael Avni wondered what would have happened if these ‘Oumuamua-like interstellar rocks were flying around 4.5 billion years ago, when our star was young and wild, and a gaseous disk was present instead of our planetary system? Their findings could be critical to answering some of the biggest puzzles regarding planet formation and the origin of planets in the Solar system  

Forming planets with foreign (planetesimal) currency

Planets are formed in protoplanetary disks, mostly made of gas and dust. The dust grains are thought to grow into pebbles, coagulate into bigger planetesimals, and finally, form planets. Once the objects reach km-size, they can survive and eventually coagulate and accrete smaller rocks/pebbles as to form planetary embryos and full-fledged planets. The main obstacle for such growth appears to occur before km-size objects form, in the stage when smaller rock and pebbles initially form. Indeed, several culprits conspire to destroy pebbles and meter-sized boulders before they can ever grow into larger planetesimals. Such pebbles and rocks move through the gaseous disk in which they are initially embedded and experience a headwind that slows them down.  The continuous push of the headwind might eventually lead them to quickly spiral inward into the Sun and be destroyed. In addition, collisions between small pebbles can lead to their fragmentation into smaller pieces halting their growth into larger planetesimals. In other words, pebbles and small rocks encounter a so-called “meter-size barrier” in their ability to grow into even larger planetesimals.    

Several models were suggested as to overcome the meter-size barrier, but these typically require fine-tuned conditions that are unlikely to exist in most planetary systems; nevertheless, it is common knowledge that most if not all stars host planetary systems. The question is then how this came to be.

In their recently published paper in the Monthly Notices of the Royal Astronomical Society, Grishin and collaborators showed that interstellar objects are the key. They suggested that most systems do not need to go through the difficult stage of forming km-size planetesimals. Instead, most systems can capture interstellar km-size planetesimals that were originally ejected from other planetary systems. But how can an object moving at tens of km per second velocity through a Solar system be captured? It turns out the answer is simple – the same headwind that drives small rocks to inspiral into their sun can slow down bigger, km-size interstellar planetesimals and thereby capture them into a newly formed protoplanetary disk. 

In this way, even a single planetary system can eject km-size planetesimals that then serve as seeds for the formation of many new planetary systems. As a result, even a very small number of planetary systems can seed the formation of many other systems – all it requires is just a few lucky rare cases to begin the process, and then these systems can spawn planetesimal “seeds” across the galaxy, which in turn can be captured into a newly forming protoplanetary disks and provide them the basic km-size building blocks needed for planetary growth. Planet formation no longer occurs in isolation; no planetary system is an island,  but rather the reservoir of ejected rogue interstellar planetesimals serves to continuously initiate the birth of new planetary systems. In turn, any newly formed planetary systems eject their own rogue planetesimals and help rebuild the reservoir of interstellar planetesimal seeds. The question becomes: what are the odds of capturing these planetesimals, and how many successful formations are required to populate the entire birth cluster with planetesimals?

Nature vs nurture: Where you live matters!

To estimate the odds of planetesimal seeding and its implications for planet formation, the researchers developed a mathematical and numerical model for capture probability, depending on the properties of the interstellar planetesimal population and the disc. They found capturing small pebbles is extremely efficient, and that capturing larger bodies is more challenging, but still reasonable.

In the dense regions of stellar clusters where tens, hundreds, or even thousands of stars are born and live in a small regions (the “Manhattan” of star formation), around 10^6 of ‘Oumuamuas are captured in the birth cluster, and the largest body captured can be as large as ~10 km.
In the galaxy’s countryside, the galactic field environment, capture is more challenging, but still around ~10^3 ‘Oumuamuas can be captured, and bodies up to ~1 km are captured per system – enough to serve as the seed for planet formation in each system. 

One is enough, planetesimals bring joy and life!

The researchers summarize that only a small fraction of the stars in a cluster (less than 1%) are required to form the primordial planetesimals, which eventually seed the entire birth cluster of ~1000 stars. Roughly similar numbers are expected also for field environments. Both estimates are conservative. The interstellar reservoir, therefore, works in tandem with the main planet formation models, providing the initial seeds for many of the planetesimal formation models.

Another interesting side aspect is that biologically active material, in the form of bacteria, can survive the tough interstellar environment if the rock in which it is embedded is large enough (larger than a few cm scale). Although only a minute fraction of ejected rocks might harbor these hardcore bacteria, a  large number of such potentially biologically active rocks can be captured. This gas-assisted capture is a far more efficient mechanism for widespread panspermia, and most systems have probably gained their first life building blocks from somewhere else.

Read all about it! New leadership, sustainability, and the scientific revolution arising around matter and light from the research labs of Prof. Motti Segev.

Students from Technion’s Viterbi Faculty of Electrical Engineering have developed an inexpensive and simple system to prevent leaving children alone in vehicles. 

The system installed in the vehicle in front of the baby safety seat.

To mark the official end of the school year, the advanced system based on machine learning technology was created by Technion undergraduate students Adam Barhak and Assaf Yitzhak under the guidance of doctoral student Ayal Taitler and master’s degree student Dotan Shambi. The system has a number of advantages that are hard to find in existing systems – ease of use, simple installation, highly reliable and the low cost. 

Leaving babies and toddlers in cars, the tragic consequences of which we are all familiar with is a human error that has occupied researchers for years. However, the various solutions offered to date, such as continuous monitoring of the weight placed on the car seat are unsatisfactory. 

Student developers of the system Adam Barhak (right) and Assaf Yitzhak.

Barhak recalls, “We asked ourselves how was it possible that no effective technological solution has been devised for this problem. We decided that we needed to change direction and embark on a new concept – an advanced and cheap thermal sensor that transfers the data to a system that is able to learn, analyze and rapidly make correct decisions.” 

The system developed by the two is based on a relatively simple and inexpensive thermal sensor installed opposite the baby seat in the back of the vehicle. The thermal sensor produces an image of the child and transfers the data to a tiny, inexpensive computer (Raspberry Pi), which processes the information and issues an alert. Additional variables are also assessed to prevent false alarms. 

The system activates a sequence of alarms in a closed loop that expands according to time passed and the temperature of the vehicle. First, a warning light is turned on, followed by a warning beep and if necessary, notification by text messages to an expanding loop of contacts. The contacts will include rescuers who can remotely open the car doors and windows. 

The system is easy to install, the sensor placed on the back of the front seat facing backward plugs into the cigarette lighter socket and can easily be transferred from one vehicle to another. The entire rear seat is monitored and an additional safety seat is not required. 

The students are gratified that their main reward for developing the system will be the drastic reduction in the number of small children abandoned in vehicles.