Lifeboat Foundation: Safeguarding Humanity
Toggle light / dark theme

Blog

Category: robotics/AI – Page 1,720

A silicon image sensor that computes

As any driver knows, accidents can happen in the blink of an eye—so when it comes to the camera system in autonomous vehicles, processing time is critical. The time that it takes for the system to snap an image and deliver the data to the microprocessor for image processing could mean the difference between avoiding an obstacle or getting into a major accident.

In-sensor , in which important features are extracted from raw data by the itself instead of the separate microprocessor, can speed up the . To date, demonstrations of in-sensor processing have been limited to emerging research materials which are, at least for now, difficult to incorporate into commercial systems.

Now, researchers from the Harvard John A. Paulson School of Engineering and Applied Sciences (SEAS) have developed the first in-sensor processor that could be integrated into commercial silicon imaging sensor chips–known as complementary metal-oxide-semiconductor (CMOS) image sensors–that are used in nearly all commercial devices that need capture visual information, including smartphones.

Existential Hope Special with Morgan Levine | On the Future of Aging

Foresight Existential Hope Group.
Program & apply to join: https://foresight.org/existential-hope/

In the Existential Hope-podcast (https://www.existentialhope.com), we invite scientists to speak about long-termism. Each month, we drop a podcast episode where we interview a visionary scientist to discuss the science and technology that can accelerate humanity towards desirable outcomes.

Xhope Special with Foresight Fellow Morgan Levine.

Morgan Levine is a ladder-rank Assistant Professor in the Department of Pathology at the Yale School of Medicine and a member of both the Yale Combined Program in Computational Biology and Bioinformatics, and the Yale Center for Research on Aging. Her work relies on an interdisciplinary approach, integrating theories and methods from statistical genetics, computational biology, and mathematical demography to develop biomarkers of aging for humans and animal models using high-dimensional omics data. As PI or co-Investigator on multiple NIH-, Foundation-, and University-funded projects, she has extensive experience using systems-level and machine learning approaches to track epigenetic, transcriptomic, and proteomic changes with aging and incorporate.
this information to develop measures of risk stratification for major chronic diseases, such as cancer and Alzheimer’s disease. Her work also involves development of systems-level outcome measures of aging, aimed at facilitating evaluation for geroprotective interventions.

Existential Hope.
A group of aligned minds who cooperate to build beautiful futures from a high-stakes time in human civilization by catalyzing knowledge around potential paths to get there and how to plug in.

Follow us!

Continue reading “Existential Hope Special with Morgan Levine | On the Future of Aging” | >

Huge breakthroughs, tiny changes: the next decade of artificial intelligence

Recent developments like DALLE-2 and LaMDA are impressive and seem ready for impact. Is AI ready to change the world?

Whether you love, fear, or have mixed feelings about the future of artificial intelligence, the cultural fixation on the subject over the past decade has made it feel like the technology’s meteoric impact is just around the corner. The problem is that it is always just around the corner, yet never seems to arrive. Many hype-filled years have passed us by since the releases of Ex Machina (2014) and Westworld (2016), but it feels like we are still waiting on AI’s big splash. However, a handful of recent developments—specifically, OpenAI’s unveiling of GPT-3 and DALLE-2, and Google’s LaMDA controversy—have unleashed a new wave of excitement—and terror—around the possibility that AI’s game-changing moment is finally here.

There are several reasons why it feels it has taken a long time for AI projects to bear fruit. One is that pop culture seems almost exclusively focused on the possible endgames of the technology, rather than its broader journey. This isn’t much of a surprise. When we stream the latest sci-fi movie or binge Black Mirror episodes, we want to see killer robots and computer chip brain implants. No one is buying a ticket to see a movie about the slow, incremental rollout of existing technology—not unless it mutates and starts killing within the first 30 minutes. But while AI’s more futuristic forms are naturally the most entertaining, and provide an endless source of material for screenwriters, anyone who based their expectations for AI off of Bladerunner has got to be feeling disappointed by now.

Deep Dive: Why 3D reconstruction may be the next tech disruptor

Were you unable to attend Transform 2022? Check out all of the summit sessions in our on-demand library now! Watch here.

Artificial intelligence (AI) systems must understand visual scenes in three dimensions to interpret the world around us. For that reason, images play an essential role in computer vision, significantly affecting quality and performance. Unlike the widely available 2D data, 3D data is rich in scale and geometry information, providing an opportunity for a better machine-environment understanding.

Data-driven 3D modeling, or 3D reconstruction, is a growing computer vision domain increasingly in demand from industries including augmented reality (AR) and virtual reality (VR). Rapid advances in implicit neural representation are also opening up exciting new possibilities for virtual reality experiences.

Swin Transformer supports 3-billion-parameter vision models that can train with higher-resolution images for greater task applicability

Early last year, our research team from the Visual Computing Group introduced Swin Transformer, a Transformer-based general-purpose computer vision architecture that for the first time beat convolutional neural networks on the important vision benchmark of COCO object detection and did so by a large margin. Convolutional neural networks (CNNs) have long been the architecture of choice for classifying images and detecting objects within them, among other key computer vision tasks. Swin Transformer offers an alternative. Leveraging the Transformer architecture’s adaptive computing capability, Swin can achieve higher accuracy. More importantly, Swin Transformer provides an opportunity to unify the architectures in computer vision and natural language processing (NLP), where the Transformer has been the dominant architecture for years and has benefited the field because of its ability to be scaled up.

So far, Swin Transformer has shown early signs of its potential as a strong backbone architecture for a variety of computer vision problems, powering the top entries of many important vision benchmarks such as COCO object detection, ADE20K semantic segmentation, and CelebA-HQ image generation. It has also been well-received by the computer vision research community, garnering the Marr Prize for best paper at the 2021 International Conference on Computer Vision (ICCV). Together with works such as CSWin, Focal Transformer, and CvT, also from teams within Microsoft, Swin is helping to demonstrate the Transformer architecture as a viable option for many vision challenges. However, we believe there’s much work ahead, and we’re on an adventurous journey to explore the full potential of Swin Transformer.

In the past few years, one of the most important discoveries in the field of NLP has been that scaling up model capacity can continually push the state of the art for various NLP tasks, and the larger the model, the better its ability to adapt to new tasks with very little or no training data. Can the same be achieved in computer vision, and if so, how?

/* */