Labeling data can be a chore. Its the main source of sustenance for computer-vision models; without it, theyd have a lot of difficulty identifying objects, people, and other important image characteristics. Yet producing just an hour of tagged and labeled data can take a whopping 800 hours of human time. Our high-fidelity understanding of the world develops as machines can better perceive and interact with our surroundings. But they need more help.

Scientists from MITs Computer Science and Artificial Intelligence Laboratory (CSAIL), Microsoft, and Cornell University have attempted to solve this problem plaguing vision models by creating STEGO, an algorithm that can jointly discover and segment objects without any human labels at all, down to the pixel.

STEGO learns something called semantic segmentation fancy speak for the process of assigning a label to every pixel in an image. Semantic segmentation is an important skill for todays computer-vision systems because images can be cluttered with objects. Even more challenging is that these objects don't always fit into literal boxes; algorithms tend to work better for discrete things like people and cars as opposed to stuff like vegetation, sky, and mashed potatoes. A previous system might simply perceive a nuanced scene of a dog playing in the park as just a dog, but by assigning every pixel of the image a label, STEGO can break the image into its main ingredients: a dog, sky, grass, and its owner.

Assigning every single pixel of the world a label is ambitious especially without any kind of feedback from humans. The majority of algorithms today get their knowledge from mounds of labeled data, which can take painstaking human-hours to source. Just imagine the excitement of labeling every pixel of 100,000 images! To discover these objects without a humans helpful guidance, STEGO looks for similar objects that appear throughout a dataset. It then associates these similar objects together to construct a consistent view of the world across all of the images it learns from.

Seeing the world

Machines that can see are crucial for a wide array of new and emerging technologies like self-driving cars and predictive modeling for medical diagnostics. Since STEGO can learn without labels, it can detect objects in many different domains, even those that humans dont yet understand fully.

If you're looking at oncological scans, the surface of planets, or high-resolution biological images, its hard to know what objects to look for without expert knowledge. In emerging domains, sometimes even human experts don't know what the right objects should be, says Mark Hamilton, a PhD student in electrical engineering and computer science at MIT, research affiliate of MIT CSAIL, software engineer at Microsoft, and lead author on a new paper about STEGO. In these types of situations where you want to design a method to operate at the boundaries of science, you can't rely on humans to figure it out before machines do.

STEGO was tested on a slew of visual domains spanning general images, driving images, and high-altitude aerial photographs. In each domain, STEGO was able to identify and segment relevant objects that were closely aligned with human judgments. STEGO's most diverse benchmark was the COCO-Stuff dataset, which is made up of diverse images from all over the world, from indoor scenes to people playing sports to trees and cows. In most cases, the previous state-of-the-art system could capture a low-resolution gist of a scene, but struggled on fine-grained details: A human was a blob, a motorcycle was captured as a person, and it couldnt recognize any geese. On the same scenes, STEGO doubled the performance of previous systems and discovered concepts like animals, buildings, people, furniture, and many others.

STEGO not only doubled the performance of prior systems on the COCO-Stuff benchmark, but made similar leaps forward in other visual domains. When applied to driverless car datasets, STEGO successfully segmented out roads, people, and street signs with much higher resolution and granularity than previous systems. On images from space, the system broke down every single square foot of the surface of the Earth into roads, vegetation, and buildings.

Connecting the pixels

STEGO which stands for Self-supervised Transformer with Energy-based Graph Optimization builds on top of the DINO algorithm, which learned about the world through 14 million images from the ImageNet database. STEGO refines the DINO backbone through a learning process that mimics our own way of stitching together pieces of the world to make meaning.

For example, you might consider two images of dogs walking in the park. Even though theyre different dogs, with different owners, in different parks, STEGO can tell (without humans) how each scenes objects relate to each other. The authors even probe STEGOs mind to see how each little, brown, furry thing in the images are similar, and likewise with other shared objects like grass and people. By connecting objects across images, STEGO builds a consistent view of the word.

The idea is that these types of algorithms can find consistent groupings in a largely automated fashion so we don't have to do that ourselves, says Hamilton. It might have taken years to understand complex visual datasets like biological imagery, but if we can avoid spending 1,000 hours combing through data and labeling it, we can find and discover new information that we might have missed. We hope this will help us understand the visual word in a more empirically grounded way.

Looking ahead

Despite its improvements, STEGO still faces certain challenges. One is that labels can be arbitrary. For example, the labels of the COCO-Stuff dataset distinguish between food-things like bananas and chicken wings, and food-stuff like grits and pasta. STEGO doesn't see much of a distinction there. In other cases, STEGO was confused by odd images like one of a banana sitting on a phone receiver where the receiver was labeled foodstuff, instead of raw material.

For upcoming work, theyre planning to explore giving STEGO a bit more flexibility than just labeling pixels into a fixed number of classes as things in the real world can sometimes be multiple things at the same time (like food, plant and fruit). The authors hope this will give the algorithm room for uncertainty, trade-offs, and more abstract thinking.

In making a general tool for understanding potentially complicated datasets, we hope that this type of an algorithm can automate the scientific process of object discovery from images. There's a lot of different domains where human labeling would be prohibitively expensive, or humans simply dont even know the specific structure, like in certain biological and astrophysical domains. We hope that future work enables application to a very broad scope of datasets. Since you don't need any human labels, we can now start to apply ML tools more broadly, says Hamilton.

STEGO is simple, elegant, and very effective. I consider unsupervised segmentation to be a benchmark for progress in image understanding, and a very difficult problem. The research community has made terrific progress in unsupervised image understanding with the adoption of transformer architectures, says Andrea Vedaldi, professor of computer vision and machine learning and a co-lead of the Visual Geometry Group at the engineering science department of the University of Oxford. This research provides perhaps the most direct and effective demonstration of this progress on unsupervised segmentation.

Hamilton wrote the paper alongside MIT CSAIL PhD student Zhoutong Zhang, Assistant Professor Bharath Hariharan of Cornell University, Associate Professor Noah Snavely of Cornell Tech, and MIT professor William T. Freeman. They will present the paper at the 2022 International Conference on Learning Representations (ICLR).

See the original post:

A new state of the art for unsupervised computer vision - MIT News

Four Syracuse University students have been awarded prestigious graduate research fellowships through the National Science Foundation (NSF) Graduate Research Fellowship Program (GRFP), and two students were recognized with Honorable Mentions. The fellowship recognizes and supports outstanding graduate students who are pursuing research-based masters and doctoral degrees in the U.S. The five-year fellowship includes three years of financial support, including an annual stipend of $34,000.

The 2022 recipients of the NSF GRFP are:

Bajorek studies medieval history and human osteology, with a research focus on the intersections of medieval medicine and religion. This fellowship will allow her to pursue her Ph.D. in medieval studies at Cornell University and provide important professional advancement resources to further her academic career. With a passion for fieldwork, Bajorek hopes to one day work as an archaeological field director at medieval cemetery sites.

As a child, I was fascinated by books about dinosaurs and Egyptian mummies, she says. An archaeological field school with Dr. Guido Pezzarossi in 2018 showed me that a career in archaeology was possible in real life. I envision that a significant portion of my career will be spent leading excavations. I eventually want to become a professor of medieval archaeology, where I can direct a teaching lab of new generations of students.

Coghiels interest in civil engineering blossomed as he was growing up in New York City, where construction projects were a constant presence. He recalls being curious about what was being developed and excited about seeing the finished products. I realized that all projects were meant to keep people safe and healthyand as I got older, I decided that I wanted to use my passion for engineering to help create a safe and sustainable future for all, he says.

With the support of the NSF GRFP, Coghiel plans to conduct research on the environmental impacts of carbon dioxide emissions from construction sites and determine whether there are more sustainable practices to implement in construction scheduling. During his undergraduate career, he participated in mentorship via the WellsLink Leadership Program and worked as a lab assistant under ECS professors Cliff Davidson and Svetoslava Todorova, both experiences he says will positively impact his career.

Working directly with engineering professors taught me numerous practices that I plan to implement in my own research project and future engineering career, Coghiel says.

Through his research, he seeks to find ways to provide healthier airspaces for all, especially underserved communities that are often adversely affected by construction pollution.

Hernndez de Lara has been interested in archaeology since volunteering with the Cuban Speleological Society in his home country of Cuba. Pursuing his Ph.D. in anthropology with a focus on historical archaeology, his research interests include conflict and battlefield archaeology and the archaeology of the contemporary past, heritage and memory. Hernndez de Lara says the NSF fellowship will support his doctoral research, and potentially open the door for future grants to continue developing his research project.

I have met great professionals at SU who have expanded my limits in unexpected ways, says Hernndez de Lara. Interacting with professors and other graduate students in the Department of Anthropology and the Maxwell School as a whole contributed to shaping my approach to the discipline, and life in general.

He hopes to ultimately become a college or university professional, with a significant interest in museums, historic preservation agencies and other public service positions in archaeology.

Tong will graduate from Syracuse next month and go on to pursue a Ph.D. in biomedical engineering at the Georgia Institute of Technology. He was drawn to biomedical engineering and a research interest in immunoengineering from an early age and aspires to have a long career working toward defeating cancer and other diseases.

The NSF fellowship provides students and their graduate institutions with a generous amount of money to support independent and creative research, giving students more flexibility to explore research directions which may not be fully supported under a faculty grant, Tong says. Now I can carefully think about what I want to do research on for my Ph.D., finding gaps in the existing research on cancer treatment options and aligning my research accordingly.

He acknowledges his work in the lab of professor Jay Henderson, as well as mentorship from professor Henderson and professor Julie Hasenwinkel, as being crucial for his success. Tongs long-term goals are to teach and mentor undergraduate and graduate students as a professor of biomedical engineering and to start his own lab to continue next-generation cancer research with curative potential.

Two students also received Honorable Mentions in this years NSF GRFP competition. Abigail McCarthy, a masters student in Earth sciences in the College of Arts and Sciences, is researching new methods to evaluate flood risk using high-throughput computing, especially regarding flood risk for socioeconomically disadvantaged communities. Karma Thomas is a Ph.D. student in biology in the College of Arts and Sciences and member of the Althoff Lab, whose research interests include the ecology and evolution of insects and their interactions with plants.

The associate provost for academic affairs and the Center for Fellowship and Scholarship Advising (CFSA) will host several faculty sessions to share ideas about how to increase the number of Syracuse University graduate and undergraduate students applying to the NSF GRFP. The sessions are as follows:

Any of the three sessions can also be joined virtually via Zoom. CFSA staff will provide a brief overview of the NSF GRFP award and the support the University currently offers to students undertaking the application. The majority of each session will be devoted to learning about faculty experiences with the program and discussing ideas to increase the number of applications.

Students interested in learning more about or applying for the next NSF GRFP award cycle or any other nationally competitive scholarships and fellowships should visit the CFSA website or email cfsa@syr.edu for more information.

Read more from the original source:

Students and Alumna Earn National Science Foundation Graduate Research Fellowships - Syracuse University News

UC San Diego is one of the biggest research institutions in the world, known for its biology, engineering, and computer science programs. Students from all over come here to study different STEM-related fields, hoping that their hard work (and the prestige of the school) will help them get high-paying and fulfilling jobs.

Looking at it from a literature major perspective, it all seems a bit daunting.

Youve probably heard the same questions that I have. A literature major? At UCSD? Butwhy? People always question my dedication to the field, my interest in it, and even my future job prospects. I once overheard a girl ask her friend, who would ever be a lit major? She was not talking about me, but I still had to resist chiming in.

The funny thing is that being a literature major was never really my plan. I applied as one because I liked writing, but I had always assumed that my final degree would be in something else. When people asked me what I planned on studying, I would say Oh, Im a literature major but thats probably going to change. I had never felt a strong connection to the subject, seeing it as more of a placeholder before I discovered what I really wanted to do.

But over time, I began to realize that graduating with a degree in literature might actually be my final goal. I really like all of the classes that I have taken so far and Ive heard that they only get better from here. There are not any other majors that I would rather pursue. Unfortunately, the more set on this path I become, the more doubts appear.

In all honesty, I do not know what Im going to do with a literature degree. My parents always told me that with a degree in the humanities, you can do everything and you can do nothing, a sentiment that is not ideal for someone as indecisive as me. No matter how much I rationalize or look up jobs or make plans for grad school, there is always this voice telling me to choose something more practical with a clear career path. Sometimes those voices are in my head, sometimes they come from other people.

I dont mind when people make fun of my major. When it comes to friends I know it is not serious, and that I can always make fun of their majors in return. But every once in a while I get reminded that people with other majors are moving so much faster than I am. STEM majors know what jobs they want when they graduate. They know how they are going to pay their bills. They are cramming for finals week when I sit off to the side with my silly little book, making my silly little highlights. And while some of these may be exaggerations or flat out false, it is how I feel.

The work you do as a literature major, or even as a humanities major, is vastly different though not necessarily easier than what you study in hard science or math. And with a school that is mostly STEM majors, that difference can feel overwhelming. I cannot help but feel that, when everyone around me is talking about clicker questions or MATLAB, I am falling behind and making the wrong decision. That once we graduate, everyone else will know what they are doing while I am still figuring things out.

I guess what I am trying to say is that being a humanities major at a majority STEM school, at least in my experience, comes with a lot of doubt. It is hard and its scary to see everyone around you doing work that is so much more tangible than yours. It is difficult to have a major that does not apply to the job market well while everyone else around you does. This also applies to people who are not a part of the humanities division, but still have a lot of doubts about their future.

What helped me was realizing that college is not a competition. I am trying not to worry about what everyone else is doing, and what everyone else is studying because my path is going to be different anyway. All of our paths are going to be different. Nobody knows what they are doing every step of the way, even though from here it seems like they might. So relax, and take a deep breath. Everything is (probably) going to be fine.

Signing off,

Samantha

Photo by Annie Spratt from Unsplash

See original here:

Dear Humanities Majors, - The UCSD Guardian Online

Funding will support childhood-to-career programs with access to digital literacy, college and career readiness, and social emotional health

Oakland, Ca Joined by Councilmembers Noel Gallo, Treva Reid and Loren Taylor, Mayor Libby Schaaf today announced Amazons support of Oakland organizations that promote a healthy and equitable culture for students and their families. The three local non-profits that will each receive $100,000 grants from Amazon are East Oakland Youth Development Center, Oakland Promise, and Tech Exchange.

If we want to change the world, we will need to do it through public and private partnerships, Oakland Mayor Libby Schaaf said. Making significant investments into our community and particularly organizations that uplift and support our youth is a true commitment to Oakland every resident can appreciate.

Working in tandem, these organizations support the varied aspects of what a child needs to be successful -- the digital tools to engage in academics and the world around them, a safe and supportive place to grow and thrive outside of school, investments in higher education opportunities, and practical skills to set them up for a productive career, said Sally Kay, Senior Manager of Amazons Public Policy and Community Engagement. Amazon celebrates these organizations because they have proven success in supporting Oakland youth and their families.

The mission of the East Oakland Youth Development Center (EOYDC) is to develop the social and leadership capacities of youth so that they achieve excellence in education, career, and service to their communities. Their mission is based on solid research which demonstrates that interventions that foster relationships with caring adults, and mastery of technical and social skills, can lead to increased well-being and civic engagement among youth from high risk environments.

Oakland Promise is available to youth and their families who reside in Oakland and offers every child the financial resources, educational programming, mentorship, and a supportive community to help them and their families to seek out and thrive in higher education and the career of their choosing.

Tech Exchange works to ensure that students and their families have a computer, Internet access, and technology skills necessary to improve and enhance their lives. Their internships allow students to work alongside technicians at schools and public housing sites to resolve technical issues, upgrade, and replace machines.

This generous investment in our organization and programming is validation of the on-going need for the safe, healing space EOYDC provides, allowing our young people to affirm their identities, explore their passions, and enhance their leadership capacities, said Landon Hill, EOYDCs Chief Program Officer. The funding that supports the work we do, is not only an investment in the young people we serve, but the community at-large who benefit from their success.

"We are honored by Amazon's investment into the youth of Oakland, said Sandra Ernst, CEO of Oakland Promise. Our city has brilliant young people that need more opportunities to pursue college and careers. Amazon's investment and partnership helps us to continue providing scholarships and program support for many of our scholars."

Since 1995, the resilient organization known as Tech Exchange has narrowed the digital divide for students and families by distributing over 75,000 computers, providing over 30,000 instructional hours of digital literacy training, and offering 200 youth technical internships, said Shinquell Green, Director of Operations for Tech Exchange. Thanks to the support of the community, our diverse team continues to bring light to and address the issue of digital equity across the Bay Area. The pandemic brought global attention to the importance of technology in every home--that it is not a luxury, but a necessity--and our mission is more critical than ever. We are grateful to Amazon for this grant, which will provide essential funding as we shift from crisis response to ongoing, sustainable digital inclusion services, tech support, and device refurbishment.

EOYDC, Oakland Promise, and Tech Exchange are all partners of #OaklandUndivided (#OU), a collective impact initiative that seeks to close the digital divide by both ensuring every Oakland public school student has home access to a computer, internet, and culturally-competent tech support and expanding access to broadband for residents across Oakland. Amazon is a multi-year Signature Partner of #OU and this recent investment enables the expansion of paid student internships to repair devices, develop IT skills, and lead awareness and outreach efforts to connect more students and families across the city.

Amazon has long been a supporter of equitable education programs, from Amazon Future Engineer, a childhood-to-career program aimed at increasing access to computer science education for children and young adults, to supporting the classroom needs of teachers through Donors Choose.

###

More:

Mayor Schaaf Announces Amazon Commitment to Support Oakland Youth - City of Oakland

Job descriptionWe're delighted to have recently won two prestigious awards the Times Higher Education UK and Ireland University of the Year 2021 and The Times and The Sunday Times Good University Guide Welsh University of the Year 2021.

If you have a record of achievement that demonstrates the commitment, aptitude and enthusiasm to contribute to our global learning community through excellent teaching, scholarship, research and innovation, along with a shared passion for our values and vision, we look forward to receiving your application. We are particularly keen to hear from, and to support, applicants who will help develop our research standing and who will enhance the diversity of our university, our community and our partnerships.

The Opportunity

Cardiff School of Technologies (CST) was formally launched in 2018 and has seen rapid growth, with now over 1,000 students engaged on a range of programmes spanning computing, digital and smart technology, data science and informatics, cyber security, electronics and robotics engineering. This rapid development has necessitated the recruitment of enthusiastic academics to contribute to the establishment and growth of the School by teaching across a range of undergraduate and postgraduate modules and in developing the underpinning research environment in key areas. Over 20 organisations have endorsed the School including Sony, the BBC, Office for National Statistics and a wide range of digital and technology companies and colleges in the Cardiff Capital Region which are work in partnership with the University to develop and deliver applied programmes.

You will be an enthusiastic and friendly professional with excellent communication and interpersonal skills coupled with a commitment to providing an excellent student experience. An established or promising research profile in the field of Technologies or related area, and teaching experience on relevant programmes is essential.

To be successful you will further enrich our students experience through your expertise and subject knowledge as evidenced through your qualifications and experience.

Principal Duties and Responsibilities

We require academic staff able to cover delivery across a range of topics such as, but not limited to:

Enthusiastic about student-centred pedagogy, you will contribute to education delivery, including programme management as required, across the range of undergraduate and postgraduate programmes. You will also make a significant contribution to employability, and professional engagement with relevant internal and external organisations, at a regional, national and international level.

You will be expected to contribute to the research profile of the department and to meet the qualitative and quantitative requirements for submission to the next Research Excellence Framework.

What we are looking for

If you would like to talk with us about this opportunity please contact Dr Paul Angel Head of Department Computer Science at PNAngel@cardiffmet.ac.uk or Stuart McNeil Head of Department Applied Computing and Engineering at SMcNeil@cardiffmet.ac.uk or for further information about working with us at our University please visit our recruitment website.

Should a high volume of applications be received, we may need to close this vacancy early. We therefore encourage you to apply as early as possible.

The University is committed to creating a highly inclusive culture. We offer family friendly and flexible working arrangements and a range of staff networks, forums and events to support and develop our people. We warmly welcome applications from those traditionally underrepresented in the higher education sector.

Further Information

Please read our Candidate Pack or visit our staff recruitment website for information about working for us, our culture, benefits, work-life balance and our career progression and development opportunities.

Go here to read the rest:

Senior Lecturer in Computer Science job with CARDIFF METROPOLITAN UNIVERSITY | 291113 - Times Higher Education