International Research Award on Computer Vision

Wednesday, May 14, 2025

The global power of artificial intelligence: Transforming healthcare and financial access

Transforming healthcare with predictive AI

“AI immediately improves response times. The second question is, can you optimise hospitals once you know that a particular area is in the red zone?

“For example, at a hospital in the US, we predicted what their emergency queue lines would look like every weekend — because people were waiting for hours in emergency queues, which sometimes resulted in deaths. By analysing previous transactional, historical, and clinical data — considering variables such as weather, payday patterns, and so on — you can actually predict quite well the kinds of ailments that are likely to appear at different times of the year.

“By doing this, we were able to improve their forecasts by 20%. As a result, they knew when to schedule more cardiologists, radiologists, nurses, and paediatric specialists — all because we were predicting the types of patients likely to come in. The emergency queue became shorter and shorter.

“This is an example of using artificial intelligence in a way that is not only beneficial for the business — in this case, the hospital — but also good for the citizens of the country. It’s about using technology to do good.”

Alternative data for financial inclusion

“68% of the adult population in emerging markets does not have access to a bank account. That means they cannot take out a mortgage, they do not have a credit card, and no one will offer them one. So, the question becomes: can alternative data, such as your digital footprint on the web, your mobile phone data, or other forms of alternative information, be used to assess creditworthiness?

“I saw an example at one of the microfinances TCO banks. There was a woman from Karachi who wanted to open a beauty salon. She had no formal education, no job, no degree — but she wanted to open a beauty salon in one of the largest megacities in the world, where the beauty and cosmetics market is growing rapidly.

“Would you give her a loan? A traditional bank probably wouldn’t. But if she had an e-wallet, you could actually assess whether she pays her bills on time, what kind of apps she has installed, what she browses online, whether she’s spending her day watching YouTube or working, studying for an accounting degree, or watching beauty tutorials. Is she making roaming calls internationally or domestically?

“You could even offer a short psychometric form — and gather insights not just from her answers, but from how quickly she types, whether she uses lowercase or uppercase letters, or a mix of both. Believe it or not, studies have shown that all of these behaviours have some correlation with credit risk profiles.

“We are doing something similar ourselves. We are currently working on a fintech product in Pakistan focused on offering nano-loans at various points along the customer journey.”

AI crisis management in public health

“We are working with one of the largest hospital networks in the US. We are also soon going to be working with the Ministry of Health in one of Asia’s largest countries, and our main focus in both cases is crisis management.

“It has become very important for governments and large hospital networks worldwide — including in Pakistan — to answer the question: how do you optimise the response a city or hospital network should have, by predicting where COVID or any other pandemic may arise next?

“As you may know, Bill Gates has already said we will likely face a pandemic every ten years or so. To become a resilient city, the Ministry of Health needs to have a dashboard. This dashboard should be connected to many different data points that can be analysed to predict where the next outbreak may occur.

“For this, we are connecting hospitals, clinics, and telecommunications data in this Asian country, pulling together information about who the people are, the areas they live in, the demographics, their symptoms, and when they had COVID.

“By doing this, we can create a heat map. For instance, if there’s an area that is highly pedestrianised, has a lot of morning traffic, or many connecting roads, we know there’s a much higher likelihood that COVID will spread there if there is a case in a neighbouring area or connected district.

“At first, this may seem obvious — and often AI initially confirms what intuition suggests — but over time, it begins to pick up on factors the human mind might miss. For example, it might detect patterns such as how many people visit a pharmacy to purchase certain medications, which could signal something underlying, like their lung health or the number of smokers in the area.“These kinds of connections can only be identified when diverse data points are combined and machine learning is applied.

Website: International Research Awards on Computer Vision #computervision #deeplearning #machinelearning #artificialintelligence #neuralnetworks, #imageprocessing #objectdetection #imagerecognition #faceRecognition #augmentedreality #robotics #techtrends #3Dvision #professor #doctor #institute #sciencefather #researchawards #machinevision #visiontechnology #smartvision #patternrecognition #imageanalysis #semanticsegmentation #visualcomputing #datascience #techinnovation #university #lecture #biomedical

Visit Our Website : computer.scifat.com Nomination Link : computer-vision-conferences.scifat.com/award-nomination Registration Link : computer-vision-conferences.scifat.com/award-registration Member Link : computer-vision-conferences.scifat.com/conference-membership/? ecategory=Membership&rcategory=Member

Awards-Winners : computer-vision-conferences.scifat.com/awards-winners

Monday, May 12, 2025

Big Trends in Embedded AI, Vision: Scaling and Multimodal Intelligence

Every year before the Embedded Vision Summit, I try to step back and reflect on the big picture in embedded AI and computer vision. This year, on the Summit’s 15th anniversary, two trends could not be clearer. First, AI and computer vision applications are moving from the lab to the real world, from science projects to widespread deployment. Second, multimodal AI—encompassing text, vision, audio and other sensory inputs—is revolutionizing what these systems are capable of.

The first trend—scaling—is wonderfully illustrated by Gérard Medioni’s keynote talk, “Real-World AI and Computer Vision Innovation at Scale.” Medioni was one of the team responsible for Amazon’s Just Walk Out cashier-less checkout technology, so he knows a thing or two about computer vision at scale. He will also discuss AI innovations that are improving the streaming experience for over 200 million Amazon Prime Video users worldwide.

Medioni’s talk will be followed by a panel discussion, “Edge AI and Vision at Scale: What’s Real, What’s Next, What’s Missing,” moderated by Sally Ward-Foxton of EE Times. On this panel, Medioni will be joined by distinguished experts from Waymo, Hayden AI,and Meta Reality Labs to discuss how vision and AI projects can go from an idea to being used by thousands or millions of people, and the challenges that must be overcome along the way.

On that same theme, Chris Padwick of Blue River Technology (a subsidiary of John Deere) will discuss “Taking Computer Vision Products from Prototype to Robust Product.” David Selinger will relate his experiences scaling up his start-up in “Deep Sentinel: Lessons Learned Building, Operating and Scaling an Edge AI Computer Vision Company.” And Jason Fayling will talk about what is needed to use AI and vision to improve operations at car dealerships in “SKAIVISION: Transforming Automotive Dealerships with Computer Vision.”

The second trend—multimodal intelligence—is spotlighted by another keynote talk, this one from Trevor Darrell of U.C. Berkeley: “The Future of Visual AI: Efficient Multimodal Intelligence.” Darrell will discuss the integration of natural language processing and computer vision through vision-language models (VLMs) and will share his perspective on the current state and trajectory of research advancing machine intelligence. Particularly relevant to edge applications, much of his work aims to overcome obstacles, such as massive memory and compute requirements, that limit the practical applications of state-of-the-art models.

Continuing the theme of multimodal intelligence, the Summit will feature several insightful talks that dive into the integration and application of multimodal AI. Mumtaz Vauhkonen from Skyworks Solutions will present “Multimodal Enterprise-Scale Applications in the Generative AI Era,” highlighting the importance of multimodal inputs in AI problem-solving. Vauhkonen will discuss the creation of quality datasets, multimodal data fusion techniques and model pipelines essential for building scalable enterprise applications, while also addressing the challenges of bringing these applications to production.

Frantz Lohier from AWS will introduce the concept of AI agents in his talk, “Introduction to Designing with AI Agents.” Lohier will explore how these autonomous components can enhance AI development through improved decision-making and multiagent collaboration, offering insights into the creation and integration of various types of AI agents. And Niyati Prajapati from Google will discuss “Vision LLMs in Multi-Agent Collaborative Systems: Architecture and Integration,” focusing on the use of vision LLMs in enhancing the capabilities and autonomy of multi-agent systems. Prajapati will provide case studies on automated quality control and warehouse robotics, illustrating the practical applications of these advanced architectures.

Because many product developers are eager to learn the practical aspects of incorporating multimodal AI into products, I will be co-presenting a three-hour training, “Vision-Language Models for Computer Vision Applications: A Hands-On Introduction,” in collaboration with Satya Mallick, the CEO of OpenCV.Org. With a focus on practical VLM techniques for real-world use cases, this class is designed for professionals looking to expand their skill set in AI-driven computer vision, particularly in systems designed for deployment at the edge.

Of course, the Summit would not be the same without its Technology Exhibits, focused on the latest building block technologies for creating products that incorporate AI and vision. The more than 65 exhibitors include Network Optix, Qualcomm, BDTI, Brainchip, Cadence, Lattice, Micron, Namuga, Sony, SqueezeBits, Synopsys, VeriSilicon, 3LC, Chips&Media, Microchip, Nextchip, Nota AI and STMicroelectronics, as well as dozens of others.

Looking back on the progress in embedded AI and computer vision over the last fifteen years, I can only shake my head in wonder. Back then, the idea of a computer being able to reliably understand images was almost science fiction. Today, machines cannot just understand images and other sensing modalities, but actually reason about them, enabling vast new classes of applications. I can barely imagine what the next fifteen years holds!

Awards-Winners : computer-vision-conferences.scifat.com/awards-winners

Sunday, May 11, 2025

The Adoption of Artificial Intelligence in Firms

Artificial intelligence (AI) could help to address sluggish productivity growth in OECD countries. This book provides evidence for policymakers, business leaders, and researchers to help understand the adoption of AI in enterprises and the policies needed to enable this. The core analysis draws on a new policy-oriented survey of AI in enterprises across the Group of Seven (G7) countries and Brazil, complemented by interviews with business representatives. The book offers a comprehensive examination of barriers to the use of AI and examines actionable solutions, including in the areas of training and education, qualification frameworks, public-private research partnerships, and public data. Also examined is the work of public institutions that seek to facilitate the diffusion of digital technologies, including AI. Further, this book highlights the need for better policy evaluation, greater international comparability in surveys of AI, and studies of generative AI in business (widespread interest in which began after this survey).

Awards-Winners : computer-vision-conferences.scifat.com/awards-winners

Saturday, May 10, 2025

AI and Machine Learning as Transformative BioTools

Nearly a century ago, Alexander Fleming discovered a penicillin-producing mold. Over the following decades, various microbes were used to make a range of therapeutics, from insulin to vaccines. Although gene-editing and other techniques can improve the production of microbe-based biologics, artificial intelligence (AI) could push these drugs even further.

“Artificial intelligence and machine learning play a crucial role as transformative tools in pharmaceutical research and microbial engineering,” according to Ayaz Belkozhayev, PhD, associate professor in the department of chemical and biochemical engineering at Satbayev University in Kazakhstan, and his colleagues. “These technologies enable the analysis of large datasets, the optimization of metabolic pathways, and the development of predictive models.”

Plus, Belkozhayev’s team points out that AI-based technologies can be used to develop efficient microbes that provide sustainable production of biotherapeutics. These biotherapeutics include ones that battle largely drug-resistant microbes, such as Acinetobacter baumannii, which can infect a person’s blood, lungs, wounds, and more.

AI-based tools could also be applied to microbes that produce lipophilic compounds, such as modified antibodies or peptides. Nonetheless, Zhang Dawei, PhD, an investigator in synthetic biology and microbial manufacturing engineering at the Tianjin Institute of Industrial Biotechnology in China, and his colleagues explain that lipophilic compounds can accumulate in cell membranes during fermentation, which can decrease production or even kill the cells producing the biotherapeutic.

To address this membrane-capturing problem, scientists explore what Dawei’s group called “membrane engineering techniques to construct highly flexible cell membranes … to break through the upper limit of lipophilic compound production.” AI could play a key role in this process. As Dawei’s group notes: “With the continuous advancement of artificial intelligence technology in the field of biomedicine, computer-assisted scientific research will provide a more comprehensive blueprint for the construction process of highly flexible cell membranes.”

Nonetheless, AI alone will not make better microbes for producing biologics. As Belkozhayev’s team emphasizes, “Innovations in genetic engineering, synthetic biology, adaptive evolution, [machine learning], and high-throughput screening have led to substantial progress in optimizing microorganisms for the efficient production of complex biological and chemical compounds.”

So, as is often the case, no one thing is the solution to all of the challenges in making biotherapeutics from microbes. Still, AI will probably enhance this area of bioprocessing.

Awards-Winners : computer-vision-conferences.scifat.com/awards-winners

Wednesday, May 7, 2025

Artificial intelligence may replace some civil servants

Artificial intelligence is now capable of performing certain tasks currently handled by government employees, according to tech entrepreneur Elon Musk. The statement was made during a closed-door session at the prestigious Milken Institute Global Conference held in Beverly Hills, California, Azernews reports.

Musk, who has long been critical of government inefficiency, reportedly said that "AI should assume many of the responsibilities currently managed by public sector workers", especially in areas burdened by bureaucracy and outdated systems. He emphasized that AI could streamline administrative tasks, improve responsiveness, and potentially save billions in taxpayer dollars.

The remarks were part of a broader conversation on technological disruption and government reform. Musk argued that modern AI systems, particularly those based on large language models, can already analyze documents, generate reports, and handle routine communications—roles traditionally assigned to human civil servants.

The session also touched on developments in Musk’s companies SpaceX and Neuralink. SpaceX continues to lead in the commercial space race, with new launches scheduled for Mars mission prototypes, while Neuralink recently announced successful trials of its brain-computer interface in human volunteers—potentially heralding a new era in medical science and human augmentation.

Although the idea of AI replacing government workers sparked debate, some attendees noted that integration rather than replacement might be the more realistic path. Others raised concerns about privacy, accountability, and job displacement in the public sector.

Awards-Winners : computer-vision-conferences.scifat.com/awards-winners

Monday, May 5, 2025

Revolutionizing Satellite Networks Through AI and Machine Learning

The satellite communications industry is undergoing a profound transformation, driven by the integration of AI/ML, cloud-native architecture, and advanced automation. These technologies are reshaping network operations by optimizing efficiency, scalability, and reliability while paving the way toward fully autonomous networks. AI/ML enables real-time anomaly detection, root cause analysis, and predictive maintenance, drastically reducing downtime and enhancing operational uptime. Cloud-native platforms deliver scalability and faster response times by processing data closer to its source, while automation ensures seamless configuration and self-healing capabilities.

Together, these innovations empower satellite networks to adapt dynamically, manage multi-orbit systems with ease, and meet growing connectivity demands. This evolution not only modernizes satellite communications but also redefines operational paradigms, heralding a new era of intelligent, self-sustaining networks capable of independent decision-making and sustained global connectivity.

Awards-Winners : computer-vision-conferences.scifat.com/awards-winners

Can deep learning transform heart failure prevention?

The ancient Greek philosopher and polymath Aristotle once concluded that the human heart is tri-chambered and that it was the single most important organ in the entire body, governing motion, sensation, and thought.

Today, we know that the human heart actually has four chambers and that the brain largely controls motion, sensation, and thought. But Aristotle was correct in observing that the heart is a vital organ, pumping blood to the rest of the body to reach other vital organs. When a life-threatening condition like heart failure strikes, the heart gradually loses the ability to supply other organs with enough blood and nutrients that enables them to function.

Researchers from MIT and Harvard Medical School recently published an open-access paper in Nature Communications Medicine, introducing a noninvasive deep learning approach that analyzes electrocardiogram (ECG) signals to accurately predict a patient’s risk of developing heart failure. In a clinical trial, the model showed results with accuracy comparable to gold-standard but more-invasive procedures, giving hope to those at risk of heart failure. The condition has recently seen a sharp increase in mortality, particularly among young adults, likely due to the growing prevalence of obesity and diabetes.

“This paper is a culmination of things I’ve talked about in other venues for several years,” says the paper’s senior author Collin Stultz, director of Harvard-MIT Program in Health Sciences and Technology and affiliate of the MIT Abdul Latif Jameel Clinic for Machine Learning in Health (Jameel Clinic). “The goal of this work is to identify those who are starting to get sick even before they have symptoms so that you can intervene early enough to prevent hospitalization.”

Of the heart’s four chambers, two are atria and two are ventricles — the right side of the heart has one atrium and one ventricle, and vice versa. In a healthy human heart, these chambers operate in a rhythmic synchrony: oxygen-poor blood flows into the heart via the right atrium. The right atrium contracts and the pressure generated pushes the blood into the right ventricle where the blood is then pumped into the lungs to be oxygenated. The oxygen-rich blood from the lungs then drains into the left atrium, which contracts, pumping the blood into the left ventricle. Another contraction follows, and the blood is ejected from the left ventricle via the aorta, flowing into veins branching out to the rest of the body.

“When the left atrial pressures become elevated, the blood drain from the lungs into the left atrium is impeded because it’s a higher-pressure system,” Stultz explains. In addition to being a professor of electrical engineering and computer science, Stultz is also a practicing cardiologist at Mass General Hospital (MGH). “The higher the pressure in the left atrium, the more pulmonary symptoms you develop — shortness of breath and so forth. Because the right side of the heart pumps blood through the pulmonary vasculature to the lungs, the elevated pressures in the left atrium translate to elevated pressures in the pulmonary vasculature.”

The current gold standard for measuring left atrial pressure is right heart catheterization (RHC), an invasive procedure that requires a thin tube (the catheter) attached to a pressure transmitter to be inserted into the right heart and pulmonary arteries. Physicians often prefer to assess risk noninvasively before resorting to RHC, by examining the patient’s weight, blood pressure, and heart rate.

But in Stultz’s view, these measures are coarse, as evidenced by the fact that one-in-four heart failure patients is readmitted to the hospital within 30 days. “What we are seeking is something that gives you information like that of an invasive device, other than a simple weight scale,” Stultz says.

In order to gather more comprehensive information on a patient’s heart condition, physicians typically use a 12-lead ECG, in which 10 adhesive patches are stuck onto the patient and linked with a machine that produces information from 12 different angles of the heart. However, 12-lead ECG machines are only accessible in clinical settings and they are also not typically used to assess heart failure risk.

Instead, what Stultz and other researchers propose is a Cardiac Hemodynamic AI monitoring System (CHAIS), a deep neural network capable of analyzing ECG data from a single lead — in other words, the patient only needs to have a single adhesive, commercially-available patch on their chest that they can wear outside of the hospital, untethered to a machine.

To compare CHAIS with the current gold standard, RHC, the researchers selected patients who were already scheduled for a catheterization and asked them to wear the patch 24 to 48 hours before the procedure, although patients were asked to remove the patch before catheterization took place. “When you get to within an hour-and-a-half [before the procedure], it’s 0.875, so it’s very, very good,” Stultz explains. “Thereby a measure from the device is equivalent and gives you the same information as if you were cathed in the next hour-and-a-half.”

“Every cardiologist understands the value of left atrial pressure measurements in characterizing cardiac function and optimizing treatment strategies for patients with heart failure,” says Aaron Aguirre SM '03, PhD '08, a cardiologist and critical care physician at MGH. “This work is important because it offers a noninvasive approach to estimating this essential clinical parameter using a widely available cardiac monitor.”

Aguirre, who completed a PhD in medical engineering and medical physics at MIT, expects that with further clinical validation, CHAIS will be useful in two key areas: first, it will aid in selecting patients who will most benefit from more invasive cardiac testing via RHC; and second, the technology could enable serial monitoring and tracking of left atrial pressure in patients with heart disease. “A noninvasive and quantitative method can help in optimizing treatment strategies in patients at home or in hospital,” Aguirre says. “I am excited to see where the MIT team takes this next.”

But the benefits aren’t just limited to patients — for patients with hard-to-manage heart failure, it becomes a challenge to keep them from being readmitted to the hospital without a permanent implant, taking up more space and more time of an already beleaguered and understaffed medical workforce.

The researchers have another ongoing clinical trial using CHAIS with MGH and Boston Medical Center that they hope to conclude soon to begin data analysis.

“In my view, the real promise of AI in health care is to provide equitable, state-of-the-art care to everyone, regardless of their socioeconomic status, background, and where they live,” Stultz says. “This work is one step towards realizing this goal.”

Awards-Winners : computer-vision-conferences.scifat.com/awards-winners

Saturday, May 3, 2025

Novel AI model inspired by neural dynamics from the brain

Researchers from MIT’s Computer Science and Artificial Intelligence Laboratory (CSAIL) have developed a novel artificial intelligence model inspired by neural oscillations in the brain, with the goal of significantly advancing how machine learning algorithms handle long sequences of data.

AI often struggles with analyzing complex information that unfolds over long periods of time, such as climate trends, biological signals, or financial data. One new type of AI model, called "state-space models," has been designed specifically to understand these sequential patterns more effectively. However, existing state-space models often face challenges — they can become unstable or require a significant amount of computational resources when processing long data sequences.

To address these issues, CSAIL researchers T. Konstantin Rusch and Daniela Rus have developed what they call “linear oscillatory state-space models” (LinOSS), which leverage principles of forced harmonic oscillators — a concept deeply rooted in physics and observed in biological neural networks. This approach provides stable, expressive, and computationally efficient predictions without overly restrictive conditions on the model parameters.

"Our goal was to capture the stability and efficiency seen in biological neural systems and translate these principles into a machine learning framework," explains Rusch. "With LinOSS, we can now reliably learn long-range interactions, even in sequences spanning hundreds of thousands of data points or more."

The LinOSS model is unique in ensuring stable prediction by requiring far less restrictive design choices than previous methods. Moreover, the researchers rigorously proved the model’s universal approximation capability, meaning it can approximate any continuous, causal function relating input and output sequences.

Empirical testing demonstrated that LinOSS consistently outperformed existing state-of-the-art models across various demanding sequence classification and forecasting tasks. Notably, LinOSS outperformed the widely-used Mamba model by nearly two times in tasks involving sequences of extreme length.

Recognized for its significance, the research was selected for an oral presentation at ICLR 2025 — an honor awarded to only the top 1 percent of submissions. The MIT researchers anticipate that the LinOSS model could significantly impact any fields that would benefit from accurate and efficient long-horizon forecasting and classification, including health-care analytics, climate science, autonomous driving, and financial forecasting.

"This work exemplifies how mathematical rigor can lead to performance breakthroughs and broad applications," Rus says. "With LinOSS, we’re providing the scientific community with a powerful tool for understanding and predicting complex systems, bridging the gap between biological inspiration and computational innovation."

The team imagines that the emergence of a new paradigm like LinOSS will be of interest to machine learning practitioners to build upon. Looking ahead, the researchers plan to apply their model to an even wider range of different data modalities. Moreover, they suggest that LinOSS could provide valuable insights into neuroscience, potentially deepening our understanding of the brain itself.

Awards-Winners : computer-vision-conferences.scifat.com/awards-winners

Tuesday, April 29, 2025

AIT – Use of artificial intelligence, big data and autonomous synthesis to accelerate the development of tomorrow’s battery technology

AIT – Use of artificial intelligence, big data and autonomous synthesis to accelerate the development of tomorrow’s battery technology

Batteries play a central role in reducing CO₂ emissions in transportation, energy, and industry. However, the development of new battery materials is still a lengthy process. It is usually based on classic trial-and-error methods that often take over a decade. To accelerate these processes, new approaches are needed that intelligently combine digital technologies and automation.

This is precisely where the European research project FULL-MAP comes in. The aim is to develop a fully integrated, AI-supported platform for accelerating material and interface development that simulates, digitalises and automates the entire battery development process – from material development to cell testing. The use of artificial intelligence, machine learning, automated synthesis and high-throughput characterization is expected to take battery development to a new technological level.

Innovation for a new generation of batteries

FULL-MAP is taking a holistic approach to drastically reducing the time from material discovery to the deployment of next-generation batteries. The key project objectives are:Building an interoperable data framework for the structured collection, use and reuse of information on battery materials and interfaces.
Development of adaptable design and simulation tools that use artificial intelligence and machine learning methods to derive suitable material structures and configurations from target specifications such as specific material properties, thereby accelerating complex simulation processes across multiple physical scales.
Further development of analysis methods and automation of high-throughput characterisation modules and technologies for fast, reliable and scalable analysis of battery materials and interfaces.
Development of AI-controlled autonomous synthesis robots to efficiently synthesise, test and further develop novel materials through data-driven iterations.
Strengthening the European research and innovation system and market positioning of the EU in the field of batteries through international cooperation and knowledge dissemination.

Awards-Winners : computer-vision-conferences.scifat.com/awards-winners

Artificial superintelligence (ASI): Sci-fi nonsense or genuine threat to humanity?

Rapid progress in artificial intelligence (AI) is prompting people to question what the fundamental limits of the technology are. Increasingly, a topic once consigned to science fiction — the notion of a superintelligent AI — is now being considered seriously by scientists and experts alike.

The idea that machines might one day match or even surpass human intelligence has a long history. But the pace of progress in AI over recent decades has given renewed urgency to the topic, particularly since the release of powerful large language models (LLMs) by companies like OpenAI, Google and Anthropic, among others.

Experts have wildly differing views on how feasible this idea of "artificial super intelligence" (ASI) is and when it might appear, but some suggest that such hyper-capable machines are just around the corner. What’s certain is that if, and when, ASI does emerge, it will have enormous implications for humanity’s future.

"I believe we would enter a new era of automated scientific discoveries, vastly accelerated economic growth, longevity, and novel entertainment experiences," Tim Rocktäschel, professor of AI at University College London and a principal scientist at Google DeepMind told Live Science, providing a personal opinion rather than Google DeepMind's official position. However, he also cautioned: "As with any significant technology in history, there is potential risk."
What is artificial superintelligence (ASI)?

Traditionally, AI research has focused on replicating specific capabilities that intelligent beings exhibit. These include things like the ability to visually analyze a scene, parse language or navigate an environment. In some of these narrow domains AI has already achieved superhuman performance, Rocktäschel said, most notably in games like go and chess.

The stretch goal for the field, however, has always been to replicate the more general form of intelligence seen in animals and humans that combines many such capabilities. This concept has gone by several names over the years, including “strong AI” or “universal AI”, but today it is most commonly called artificial general intelligence (AGI).

"For a long time, AGI has been a far away north star for AI research," Rocktäschel said. "However, with the advent of foundation models [another term for LLMs] we now have AI that can pass a broad range of university entrance exams and participate in international math and coding competitions."

Awards-Winners : computer-vision-conferences.scifat.com/awards-winners

Saturday, April 26, 2025

Artificial intelligence models fall short of predicting social interactions

A study from Johns Hopkins University reveals that humans excel over AI in understanding social interactions in motion, an essential skill for technologies such as self-driving cars and assistive robots.

Current AI struggles to recognize human intentions, such as whether a pedestrian is about to cross the street or if two people are engaged in conversation. Researchers suggest that this issue stems from how AI is built, as it cannot fully grasp social dynamics.

To compare AI models with human perception, researchers had people watch short video clips and rate how well they understood the social interactions depicted. The videos showed people engaging with each other, doing activities side by side, or acting independently.

Next, they tested over 350 AI models—spanning language, video, and image processing—asking them to predict how humans would judge the videos and how their brains might respond. For large language models, they analyzed human-written captions to see how well AI understood social dynamics.

Researchers found that human participants generally agreed on how they interpreted social interactions in videos, but AI models struggled, regardless of their training data or size.

This highlights a major gap in AI’s understanding of unfolding social dynamics. Scientists believe this limitation stems from AI’s design, as current models are built like the part of the human brain that processes static images, unlike the region responsible for interpreting dynamic social scenes. The study suggests that AI cannot still fully mimic how humans naturally perceive and respond to social interactions.

Awards-Winners : computer-vision-conferences.scifat.com/awards-winners