Uber’s robots will be strategically positioned in busy areas with restaurants and crowds, aiming to bring a festive atmosphere to these demanding spaces.
Uber’s holiday season activities also include a festive fleet, Christmas kit, and Gift hub.
Researchers are utilizing a reinforcement learning method known as ‘curiosity-driven’ training for tasks like calling the elevator, opening doors, and sorting boxes.
Swiss-Mile’s ANYmal robot is being trained to perform practical tasks by manipulating its wheeled legs and arms.
The robots offer a modular, multi-functional tool may help to efficiently and economically execute a spectrum of tasks on the lunar surface.
The Japanese startup was selected as part of 14 firms selected by DARPA for its LunA-10 study.
The model expects to generate $203 million in revenue by 2030 from its new health promotion business.
How would you feel if someone offered you a stay at a hotel, but with a catch: they would watch you sleep and collect data from your body? Would you be curious, excited, or creeped out? Well, that’s exactly what a Japanese IT company plans to do as part of its new health promotion business.
NTT’s bold plan
Continue reading “Why a Japan firm plans to track sleep pattern of 10 million hotel guests” »
The alliance aims to open-source the development of artificial intelligence and take on the bad boys of AI, Microsoft, OpenAI, and Google.
Major names in the technology industry, such as IBM, Meta, and many others who seemed to have been left out of the race to develop artificial intelligence (AI) models, have now teamed up to form the AI Alliance.
The collaborative effort also includes government and research organizations and a few startups that will work together to “support open innovation and open science in AI”, a press release from IBM about the alliance said.
Continue reading “IBM, Meta lead 50+ tech firms to counter AI dominance of OpenAI, Google” »
Summary: Mice display behavior akin to self-recognition when viewing their reflections in mirrors. This behavior emerges under specific conditions: familiarity with mirrors, socialization with similar-looking mice, and visible markings on their fur.
The study also identifies a subset of neurons in the hippocampus that are crucial for this self-recognition-like behavior. These findings provide valuable insights into the neural mechanisms behind self-recognition, a previously enigmatic aspect of neurobehavioral research.
Diverse and full of sea life, the Earth’s Devonian era—taking place more than 370 million years ago—saw the emergence of the first seed-bearing plants, which spread as large forests across the continents of Gondwana and Laurussia.
However, a mass extinction event near the end of this era has long been the subject of debate. Some scientists argue the Late Devonian mass extinction was caused by large-scale volcanic eruptions, causing global cooling. Others argue a mass deoxygenation event caused by the expansion of land plants was to blame.
A recently published study in the journal Communications Earth & Environment led by researchers at IUPUI now posits that both factors played a role—and draws attention to the environmental tipping points the planet faces today.
To achieve high intrinsic gain (Ai) in OTFTs, it is necessary to enlarge output resistance (ro) or transconductance (gm) according to a typical formula of Ai = gmro, which is very difficult for conventional OTFTs because of inherent device structure and operating mode limitations (11, 12). Recently, the “Schottky barrier” (SB) strategy based on metal-semiconductor junction (MS junction) has been adopted in TFTs to pursue high-gain and low-saturation voltage, including subthreshold SB-TFTs (11, 12, 15, 16) and source-gated transistors (17, 18). Unfortunately, the subthreshold transistors are limited in low and narrow subthreshold operating region rather than the normal ON-state region (namely, the normal voltage operating region in a typical TFT), which are difficult to be compatible with typical circuits. As far as we know, the ultrahigh-gain (1000) OTFTs operating in the ON-state region have not been previously reported. On the other hand, the state-of-the-art OTFTs above have mostly suffered from uncontrollable barriers owing to energy-level mismatches and a series of complex interface problems, such as Fermi-level pinning and interface chemical disorder (19). In this case, considerable low-energy carriers are allowed to pass through the junction by thermionic field emission and tunneling models instead of thermionic emission model, which is not conducive to obtaining a high output resistance and high intrinsic gain. Most barrier heights in MS junction do not conform to the prediction value of Schottky-Mott rule. Theoretically, an ideal and high-quality barrier with thermionic emission model allows the rapid depletion of carriers at the source electrode, thus yielding ultrahigh gain, infinite output resistance, and low saturation voltage (11, 12). In addition, infinite output resistance at the saturation regime indicates that the output current is very stable and flat. This performance is helpful because only a single OTFT is used as a simplified current stabilizer in circuits without complex circuit design, which benefits low power and low cost in circuits. Therefore, it is necessary to develop a high-quality barrier strategy to modulate charge injection to meet the requirements of ultrahigh-gain OTFTs.
Here, we demonstrate a metal-barrier interlayer-semiconductor (MBIS) junction to prepare high-performance MBIS-OTFT with an ultrahigh gain of ~104 in the ON-state region, low saturation voltage, almost negligible hysteresis, and good stability. On the basis of low-energy processes and in situ surface oxidation technology, the high-quality van der Waals MBIS junction with wide-bandgap semiconductor (mainly Ga2O3) interlayer is achieved, allowing for an adjustable barrier height and thermionic emission properties. A series of in situ experiments and simulations revealed the relationship between the barriers and the device’s performance. Furthermore, as demonstrations, a simplified current stabilizer and an ultrahigh-gain organic inverter are exhibited without complex circuit design.
A team of AI researchers at Google’s DeepMind project have developed a type of AI system that is able to demonstrate social learning capabilities. In their paper published in the journal Nature Communications, the group describes how they developed an AI application that showed it was capable of learning new skills in a virtual world by copying the actions of an implanted “expert.”
Most AI systems, such as ChatGPT, gain their knowledge through exposure to huge amounts of data, such as from repositories on the Internet. But such an approach, those in the industry have noted, is not very efficient. Therefore many in the field continue to look for other ways to teach AI systems to learn.
One of the most popular approaches used by researchers is to attempt to mimic the process by which humans learn. Like traditional AI apps, humans learn by exposure to known elements in an environment and by following the examples of others who know what they are doing. But unlike AI apps, humans pick things up without the need for huge numbers of examples. A child can learn to play the game of Jacks, for example, after watching others play for just a few minutes—an example of cultural transmission. In this new effort, the research team has attempted to replicate this process using AI constrained to a virtual world.
Using single-cell and spatial transcriptomics, human embryonic limb development across space and time and the diversification and cross-species conservation of cells are demonstrated.
