Make no mistake about it, there is a lot of money being spent on generative AI in 2025.
Category: robotics/AI – Page 392
Revolutionizing 3D vision: How miniaturized snapshot polarization imaging is transforming depth sensing
Capturing precise 3D details with a single camera has long been a challenge. Traditional methods often require complex dual-camera setups or specialized lighting conditions that are impractical for real-world applications. However, a groundbreaking approach developed at Nanjing University is set to redefine 3D imaging.
In our latest research, published in Optica, we introduce a cutting-edge snapshot polarization stereo imaging system (SPSIM), as shown in Fig. 1. This innovative system integrates metasurface optics with artificial intelligence to extract highly detailed 3D shape information in real time.
Unlike conventional methods that rely on multiple polarizers or sequential exposures, SPSIM utilizes a specially engineered metasurface lens to capture full-Stokes polarization data in a single shot. With an extinction ratio of 25 dB—comparable to commercial polarizers—and an unprecedented central wavelength efficiency of 65%, our system outperforms standard polarization cameras.
World’s smallest battery-free flying robot uses magnets for power
UC Berkeley’s tiny, bee-inspired flying robot weighs 21 mg, hovers, and hits targets using magnetic fields for potential rescue missions.
China’s KEENON unveils smart humanoid robot for hospitality sector
Keenon Robotics unveils XMAN-R1 humanoid and new KLEENBOT robots, advancing automation in hospitality and service industries.
Homeowners share honest review after living in futuristic house built by robots: ‘We are a part of the future’
Everything is bigger in Texas — including a new housing development with a futuristic vision.
Icon, a 3D technology company, is behind dozens of next-generation 3D-printed homes in the Lone Star State. A YouTube video gives viewers an inside look at the new homes built with robotic construction at the Wolf Ranch development in Georgetown.
Emergence AI’s new system automatically creates AI agents rapidly in realtime based on the work at hand
Hailed by various market research reports as the big tech trend in 2025 — especially in the enterprise — it seems we can’t go more than 12 hours or so without the debut of another way to make, orchestrate (link together), or otherwise optimize purpose-built AI tools and workflows designed to handle routine white collar work.
Yet Emergence AI, a startup founded by former IBM Research veterans and which late last year debuted its own, cross-platform AI agent orchestration framework, is out with something novel from all the rest: a new AI agent creation platform that lets the human user specify what work they are trying to accomplish via text prompts, and then turns it over to AI models to create the agents they believe are necessary to accomplish said work.
This new system is literally a no code, natural language, AI-powered multi-agent builder, and it works in real time. Emergence AI describes it as a milestone in recursive intelligence, aims to simplify and accelerate complex data workflows for enterprise users.
Anyone can run quantum simulations thanks to new chatbot for chemistry
At times, the reactions do not produce the intended results, and this is where simulations are used to understand what might have caused the anomalous behavior. Chemistry students are often tasked with running these simulations to learn to think critically and make sense of discoveries.
As the complexity of the process increases, more advanced computing infrastructure is required to carry out these simulations. To understand these reactions at a quantum level, theoretical chemists even use specialized software packages to streamline their research and automate the simulation process. AutoSolvateWeb is just a chatbot but can help even non-experts achieve this level of competence.
AutoSolvateWeb helps compute the dissolving of a chemical, referred to as a solute, into a substance called a solvent. The resultant solution is called the solvate, hence the name. While theoretical chemists use computation software to convert this into simulations that look much like 3D movies, AutoSolvateWeb can achieve the same output through a chatbot-like interface with the user.
AI thinks like us—flaws and all: Study finds ChatGPT mirrors human decision biases in half the tests
Can we really trust AI to make better decisions than humans? A new study says … not always. Researchers have discovered that OpenAI’s ChatGPT, one of the most advanced and popular AI models, makes the same kinds of decision-making mistakes as humans in some situations—showing biases like overconfidence of hot-hand (gambler’s) fallacy—yet acting inhuman in others (e.g., not suffering from base-rate neglect or sunk cost fallacies).
Published in the Manufacturing & Service Operations Management journal, the study reveals that ChatGPT doesn’t just crunch numbers—it “thinks” in ways eerily similar to humans, including mental shortcuts and blind spots. These biases remain rather stable across different business situations but may change as AI evolves from one version to the next.
How neural networks represent data: A potential unifying theory for key deep learning phenomena
How do neural networks work? It’s a question that can confuse novices and experts alike. A team from MIT’s Computer Science and Artificial Intelligence Lab (CSAIL) says that understanding these representations, as well as how they inform the ways that neural networks learn from data, is crucial for improving the interpretability, efficiency, and generalizability of deep learning models.
With that mind, the CSAIL researchers have developed a new framework for understanding how representations form in neural networks. Their Canonical Representation Hypothesis (CRH) posits that, during training, neural networks inherently align their latent representations, weights, and neuron gradients within each layer. This alignment implies that neural networks naturally learn compact representations based on the degree and modes of deviation from the CRH.
Senior author Tomaso Poggio says that, by understanding and leveraging this alignment, engineers can potentially design networks that are more efficient and easier to understand. The research is posted to the arXiv preprint server.