Cloud Hosting

Ive spent two decades as a researcher and educator in artificial intelligence, drawn to the field by the opportunity to explore the mysteries of perception and cognition. But life is rarely as simple as wed like, and the arc of my career has paralleled my mothers escalating health struggles, including a chronic, life-threatening cardiovascular condition. As all-consuming as the world of academia can be, it sometimes feels as if Ive spent as much time in hospitals as I have in my lab.

Im happy to report my mother continues to persevere, but her resilience hasnt been the only silver lining to this ordeal. Years spent in the company of nurses and doctorsunfailingly committed, but perpetually overworked and often sleep deprivedconvinced me that the power of AI could radically elevate the way care is delivered. Intelligent sensors could keep tireless watch over patients, automate time-consuming tasks like charting and transcription, and identify lapses in safety protocols as they happen. After all, if AI can safely guide cars along freeways at 70 miles per hour, I wondered, why cant it help caregivers keep up with the chaos of the healthcare environment?

At the heart of this idea was an obstacle, however. I was proposing research that extended beyond the limits of computer science and into an entirely different field, with decades of literature and traditions stretching back generations. It was clear I needed a collaboratornot just an authority in healthcare, but one with the patience and open-mindedness to help an outsider bring something new to the table. For the first time in my career, success would depend on more than the merits of my work; it would require the humility of researchers like me to recognize the boundaries of our knowledge, and the graciousness of experts in another discipline to help us overcome them.

Thankfully, luck was on my side. In 2012, a colleague introduced me to Arnie Milstein, a Stanford Medical School professor and member of the National Academy of Medicine with an interest in both the policy and the technology that drives healthcare. Our first real conversation on the topic turned a casual lunch at a Vietnamese pho restaurant into an impromptu, hours-long brainstorming session. The exuberance of that day never wore off, as we convened a coalition of researchers to explore the automated tracking of surgical tools during operations, privacy-preserving monitors that ensure the safety of high-risk patients and vulnerable seniors, and networks of smart sensors that help hospital staff maintain hand hygiene throughout their shifts. Finally, in September, after years of experimentation, refinements, and presentations at conferences all over the world, our research was published in Nature. And now, with the help of legal scholars, bioethicists, and even a philosopher, were partnering with select hospitals and senior homes to pilot its use in the hands of real caregivers.

The success of my collaboration with Professor Milstein demonstrates an important idea: AIs applications are vast, but technology will represent only part of any given breakthrough. The remainder will be found in the contributionseven leadershipof experts from a growing list of fields, of which healthcare is only one example. Similar partnerships await as AI intersects with economics, energy, environmental science, public health, education, and even the humanities.

For instance, its hard to talk about any application of technology in 2020 without addressing the coronavirus pandemic. This was among the motivating factors behind the launch of AI Cures, an MIT initiative that brings together researchers in machine learning and life sciences to accelerate the speed with which antivirals can be identified, evaluated, and ultimately deployed. Its applications in the face of COVID-19 are obvious, but its broader goal of elevating our defense against pathogens of all kinds will remain relevant long after the challenges of the present moment are behind us. In addition to its core research mission, the group has organized impressively inclusive events in recent months, providing a venue for presenters with backgrounds in computer science, infectious disease, cardiology, synthetic biology, and many others.

Similarly encouraging is the work of my colleague, Stanford law professor Dan Ho. His lab has published extensively on the utility of AI in the public sector, and is now working with the EPA to use machine learning to dramatically improve the tracking of ecological contamination at a national scale. The underlying technology is transformative, but its the involvement of legal scholars, policymakers, and government representatives that truly makes it applicable in the real world.

These stories are a testament to the power of humility, but the sheer scale of the challenges that remain calls for a more organized response. It was with this in mind that I partnered with Stanford professor of philosophy and former provost John Etchemendy to co-found the Stanford Institute for Human-Centered Artificial Intelligence, or HAI, in 2018. Its ongoing mission is to reframe the pursuit of AI in unequivocally human terms, to reflect its dependence on interdisciplinary alliances, and to ensure ethics, compassion, and societal responsibility are baked in from the earliest stages of our workwhether its an algorithm, a commercial product, or even legislation.

HAIs reach as an institution is helping to cross new divides as well, beyond those that separate academic worlds. Partnerships with corporations, governments, and NGOs, for instance, will be essential in building a larger community around these values. Already, for example, theyve helped us organize cross-disciplinary workshops that bring ethical, philosophical, and legal expertise to bear on contentious technologies like facial recognition, with audiences of executives and legislators at both the state and federal level. And our relationships with tech leaders like Google and Amazon allow us to offer powerful cloud computing accessa foundational but often cost-prohibitive resource for modern AI researchto young, innovative thinkers in the form of grants.

Ultimately, however, this appreciation for the power of humilityopenness, transparency, and a reverence for the expertise of otherscant be mandated from the top down. It must be built up from a cultural level, and thus requires an investment in educational efforts to instill them in the next generation of AI practitioners. Here at Stanford, political science professor Rob Reich co-created a course in the computer science department entitled Computers, Ethics and Public Policy, intended to augment the education of engineers with an awareness of their impact on people and communities, while Harvard computer science professor Barbara Grosz explores similar issues in a course called Embedded Ethics. These are encouraging signs of a shift in the way we educate not just tomorrows technologists, but business leaders, social scientists, and politicians. Its my hope that universities across the world will be inspired to follow suit.

The excitement and anxiety surrounding AI can lend it a fatalistic tone, with aggressive language like revolution, tectonic shift, and force for change all too common. But while it might seem inevitable that AI will reshape the world, collaborations like these are a chance for the worldin all its messy, complicated vibrancyto reshape AI in turn. So although Im continually excited by what were learning about intelligent machines, Im even more excited by what we can learn from each other. All it takes is the willingness to ask, and that great, understated strengthour humility.

Dr. Fei-Fei Li is the Sequoia Professor, Computer Science Department, and Denning Codirector, Stanford Institute for Human-Centered Artificial Intelligence, Stanford University. She is an elected Member of the National Academy of Engineering, and the National Academy of Medicine.

Go here to see the original:

Stanford AI scholar Fei-Fei Li writes about humility in tech - Fast Company

Many traffic lights today are managed by simple computers called signal controllers.

Getty Images

Traffic lights at intersections are managed by simple computers that assign the right of way to the nonconflicting direction. However, studies looking at travel times in urban areas have shown that delays caused by intersections make up 12-55% of daily commute travel, which could be reduced if the operation of these controllers were more efficient.

A team of researchers led by Guni Sharon, professor in the Department of Computer Science and Engineering at Texas A&M University, has developed a self-learning system that uses machine learning to improve the coordination of vehicles passing through intersections.

The researchers published their findings in the proceedings of the2020 International Conference on Autonomous Agents and Multiagent Systems.

Many traffic signals today are equipped with signal controllers that serve as the brains of an intersection. They are programmed with various settings to tell the traffic display when to change colors depending on the time of day and traffic movement. This gives the signals the ability to handle fluctuations in traffic throughout the day to minimize traffic congestion.

Recent studies have shown learning algorithms, based on a concept in psychology called reinforcement learning where favorable outcomes are rewarded, can be used to optimize the controllers signal. This strategy enables controllers to make a series of decisions and learn what actions improve its operation in the real world. In this instance, the result would be a reduction in the buildup of traffic delays.

But Sharon noted that these optimized controllers would not be practical in the real world because the underlying operation that controls how it processes data uses deep neural networks (DNNs), which is a type of machine-learning algorithm. They are commonly used to train and generalize a controllers actuation policy, which is the decision-making (or control) function that determines what actions it should take next based on the current situation its in. It consists of several sensors that give information about the current state of the intersection.

Despite how powerful they are, DNNs are very unpredictable and inconsistent in their decision-making. Trying to understand why they take certain actions as opposed to others is a cumbersome process for traffic engineers, which in turn makes them difficult to regulate and understand the different policies.

To overcome this, Sharon and his team defined and validated an approach that can successfully train a DNN in real time while transferring what it has learned from observing the real world to a different control function that is able to be understood and regulated by engineers.

Using a simulation of a real intersection, the team found that their approach was particularly effective in optimizing their interpretable controller, resulting in up to a 19.4% reduction in vehicle delay in comparison to commonly deployed signal controllers.

Despite the effectiveness of their approach, the researchers observed that when they began to train the controller, it took about two days for it to understand what actions actually helped with mitigating traffic congestion from all directions.

Our future work will examine techniques for jump starting the controllers learning process by observing the operation of a currently deployed controller while guaranteeing a baseline level of performance and learning from that, Sharon said.

Other contributors to this research include Josiah P. Hanna, research associate in the School of Informatics at the University of Edinburgh, and James Ault, doctoral student in the Pi Star Lab at Texas A&M.

View original post here:

Optimizing Traffic Signals To Reduce Intersection Wait Times - Texas A&M University Today

Juniors and Quincy House residents Zeel Patel and Eric Lin excitedly took the reins of the Harvard College Project for Asian and International Relations last summer as its new co- presidents. They looked forward to hosting the groups two flagship student-run conferences one in Cambridge and one in an Asian city carrying on the 30-year-old groups longtime tradition of creating spaces for a global cohort of students and young professionals to connect to one another.

But when the COVID-19 pandemic worsened, they had to cancel their summer conference in Taipei, scheduled for August. They were disappointed, but undeterred. Given the logistical obstacles and ongoing political and public health uncertainties, Patel and Lin designed a virtual version of their winter conference, which would normally take place on Harvards campus.

This weekend, young Asian professionals, students, and others from around the country will join them to learn from speakers including Tedros Adhanom, director-general of the World Health Organization; Kevin Sneader, the CEO of consulting firm McKinsey & Company; and Yang Lan, a media executive and talk-show host known as the Oprah Winfrey of China. So far, more than 1,800 participants have registered to attend.

We realized a lot of the friction points for running a conference were no longer there in a virtual setting. Speakers wouldnt have to fly in to our site and stay in hotel rooms, and then speak at the conference. They could just dial in from anywhere, and the same [was true] with the delegates, said Patel, a computer science and neuroscience concentrator from Calgary, Canada. So we wanted to be really ambitious about the speakers we invited and how we organized it.

Using their computer science skills and knowledge of the Silicon Valley start-up world, Patel and Lin secured a contract with events platform Run the World to manage delegate coordination, breakout rooms, one-on-one pairing with mentors, and social events. Its a move that the team hopes will help bring the intimacy and collaborative spirit of an in-person gathering to delegates logging on from more than 10 time zones worldwide.

The theme of the conference is Embracing Change, with a focus on the democratization of information and providing networking resources to students and young people who lack the access to high-powered individuals that a Harvard education can provide. The conference is free for Harvard undergraduates and HPAIR offers scholarships for low-income delegates, no matter where they are located.

Eric and I saw this as a really cool opportunity to leverage the power of software and make our conference super-accessible to everyone, said Patel.

We see these resources offered at well-off institutions like Harvard, and in Western countries more generally, but the same opportunities are not present in other areas, added Lin, a computer science concentrator from McLean, Virginia. We want delegates to see what, for example, the CEO of McKinseys vision is for the world, and be able to talk to him in an intimate setting. I hope it triggers inspiration and ignites some kind of motivation for them to explore beyond their local communities.

The changes to the conference format havent deterred interest from would-be participants; sign-ups from applicants for the 2021 meeting are more than three times what they have been in the past. Lin attributed the increase to the outreach and promotion work by the planning team of 40 Harvard students and their 30 counterparts at Asian universities.

The growth of the conference also allowed the team to expand in a new way. In previous years, financial margins were tight. But after the unexpected bump in revenue and reduced costs from this years conference, HPAIR will donate a portion of the funds raised by the conference to the Bill and Melinda Gates Foundation.

These achievements are a testament to how hard our team has worked to get this up and running and to grow it in a completely different new way, [and was] a full team effort, said Lin.

I hope Harvard students [who attend] can see the different cool organizations that are part of Harvards campus and what they can be involved in when theyre back on campus, said Patel. [Eric and I] have similar immigrant backgrounds, and [with HPAIR], its been exciting to see what life is like in developing countries, and then put a spotlight on it so that other people who havent had that exposure can see it.

HPAIR Harvard Conference 2021: Embracing Change runs Jan. 15-18. Content from the conference will be available on the groups YouTube channel later this month.

View post:

Student-run HPAIR conference goes virtual this year - Harvard Gazette