|
Matan Ninio, IBM Researcher |
Matan Ninio, an expert in machine-based
anomaly detection at IBM Research - Haifa,
is also a born hacker, a maker, a computational biologist, t-shirt designer,
and a devoted family man. Matan believes we are privileged to live during the
3D printing revolution, which is allowing makers to depart from mass
manufacturing and enter a new world of customizing solutions for special
individual needs. He’s proud to be part of this revolution and to bring new
tools to the people who really need them.
I recently spoke with him about his work
and his volunteer ‘maker’ contributions to TOM,
short for Tikun Olam, a Hebrew term that means making the world a better place.
Where did you grow up?
Matan Ninio: I grew up in Jerusalem, but also lived in
the US and England for a while. Having a mother who is an academic and
professor of psychology meant frequent travel for sabbaticals. I got my
bachelors as part of my service in the Israeli army’s Talpiot program and
then continued on to do a masters and PhD in computational biology and machine
learning at the Hebrew University.
When did you join IBM, and
why?
MN: I joined IBM five years ago. The machine
learning group in Haifa has built a strong reputation as a preferred place for
machine learning students, especially at Hebrew U. Many colleagues who left the
university came here, so it was a matter of rejoining old friends. I knew there
were good people here.
What is your area of
expertise?
MN: For the past few years I’ve been working
on a Research project called zAware,
an integrated solution that helps identify unusual system behavior in real
time. The technology was recently contributed to the open source community for Linux,
which was a nice testimony to its importance for IT systems.
We listen to the data from the mainframe
logs and use machine learning to detect when the machine’s behavior changes to
something abnormal. On a stable product machine, this is usually an indication
that something is going wrong. The power of our approach lies in finding things
that are unpredictable and not necessarily actions that are flagged as errors.
How does your Maker community
involvement matter?
MN: My machine learning group at IBM Research
– Haifa has been building a name for itself as a fast response specialist.
We’re winning hackathons in different domains, and this is where it started.
|
Matan Ninio (R) and user of the prosthetic hand with sensors
(credit: ZOA production and TOM) |
One interesting initiative coming out of
the Israeli maker community is TOM. The organization brings together the maker
community and the special needs community to design and create solutions that
require customization, or are produced only in small quantities. The organization gets makers and need-knowers (people who understand the needs) together for
three day hackathons, with enough tools and experts to build things that both
look good and function well.
One of the most interesting maker
contributions is the design for a prosthetic hand. Several years ago, a
carpenter who had lost three fingers, and a special effects expert from
Vancouver designed a hand for everyday use. The hand could be printed on a 3D
printer with a few wires and springs for about $50 a piece – as an alternative
to prosthetics that cost thousands of dollars per finger.
We set out to add sensorial abilities to
these hands so, for example, the person could lift a paper cup full of water
using just the right amount of pressure. There are a lot of projects to build
hands that use sensors to activate the fingers for typing, holding a cup, or
playing ball. But there’s no way for the person to monitor how much pressure
they’re applying.
To get an idea of the problem we set out
to tackle, imagine trying to adjust a car window to hold a paper cup between
the window and the car frame using only the up and down buttons. Getting the
pressure and movement right so you don’t crush the cup is almost impossible. Our
goal was to give sensory feedback on the amount of force used. We developed a
way of feeding the information back to the person using pressure applied to the
forearm, in an area where the arm naturally feels the muscle working. To turn
this from a working idea into something in the field is not trivial but we are
continuing to work on it. It was really
nice to be awarded a prize for leading the effort to build an intelligent hand
prosthesis.
|
Matan (R), with another TOM volunteer, testing the box
for hydrotherapy (Collaboration with Beit Issie Shapiro) |
In another TOM challenge, we enabled eye
tracking devices to work in wet areas. We worked with a young girl who suffers from Rett syndrome. She has only minimal physical abilities and
communicates via eye movement by looking at onscreen buttons which then speak
out the button’s word.
Hydrotherapy is very helpful for the girl's condition,
but she and others like her have no way to communicate with their
therapist while in the water. In this case, we actually thought inside the box.
We built a waterproof box that protects the computer device, but also ensures
venting of excess heat so the computer doesn’t overheat. We added a Bluetooth
water mouse, speakers, and other tech so the system could be used at the pool.
So far, the responses and feedback are
very good. And we are working on sharing the instructions on how to put this
together so other people around the world can use it. There is a real need.
What's your advice to someone
who wants to be a "maker"?
MN: I feel I’m really making a difference and
influencing the lives of people who need help. It’s one of the most encouraging
things I do. We are at an intersection in time where technologies like 3D
printing are taking the creation of tools and equipment away from the factories
and giving makers the ability to build custom solutions to meet the needs of
individuals. I’m proud to be part of this revolution and impact the lives of
people. And, of course, it’s great fun to create things.
Labels: 3D printing, IBM Research - Haifa, machine learning, maker, Matan Ninio, TOM