Xiaomi establishes R&D center in Finland for developing smartphone intelligent camera technology.
2019-10-31
Xiaomi fulfills promise by opening camera technology research center in Tampere, Finland, leveraging Nokia's past expertise. The focus is on machine learning, camera algorithms, and intelligent image processing to advance smartphone camera technology. Huawei's prior establishment in Finland adds to competition with Xiaomi in camera technology innovation.
Read more ›AI Teaches Harvesting Robots to Assemble Objects
2019-11-01
Researchers study Form2Fit, a machine learning system for teaching pick-and-place robots to assemble objects using suction arms. Training dataset created through time reversal disassembly. Robot achieves 94% success in various toolkit configurations. Researchers aim to enhance Form2Fit's practicality.
Read more ›Google launches new features for their AI platform.
2019-10-28
Google's AI Platform now supports Nvidia graphics chips for training on any hardware. It offers data labeling, Jupyter Notebook, and Prediction with new backend for online predictions. AI Platform Training allows running scripts on various hardware and supports custom frameworks for training models in any language. Cloud AI Platform simplifies model training and deployment for solving challenging problems.
Read more ›Epic Challenge of "What to Wear Today?" Easily Solved by Artificial Intelligence...
2019-10-28
Fashion++ is a Facebook app using AI for personalized fashion suggestions. It refines existing outfits to make them more stylish, maximizing wardrobe use without excessive spending. The app utilizes a neural network and editing modules to generate outfit modifications. Fashion++ has been well received for its fashionable and easy-to-implement suggestions. On the other hand, "WANNA KICKS" is an AR app allowing users to virtually try on shoes with smartphone cameras, offering a seamless and realis
Read more ›The Fluctuations in the Field of Robotics
2019-10-28
The robotics industry has its ups and downs, with many companies facing financial challenges. ABB invests $150 million in a Shanghai factory, aiming to lead in Industry 4.0. Meanwhile, Boston Dynamics showcases advanced robots, although the "killer" robots are just a special effect. Unfortunately, Carbon Robotics, known for low-cost robotic arms, has gone bankrupt due to funding issues. Commercialization is key in the robotics sector, with success leading to growth and failure resulting in bankr
Read more ›Google Releases Latest SGD New Corpus to Help Smart Assistants Become Smarter
2019-10-29
Google researchers have developed a method to enhance AI assistant services without retraining, introducing the SGD corpus with 18,000+ annotated dialogues across 17 domains. They emphasize the importance of adaptability and the need for large-scale datasets to improve virtual assistant performance. Unified models based on natural language descriptions can facilitate dialogue systems and enable operation on new services. The SGD corpus is a valuable resource for building dialogue models, advanci
Read more ›Enhancing Safety Systems for Autonomous Vehicles
2019-10-29
MIT engineers have developed a system to enhance safety in autonomous driving by perceiving ground shadows. This may help avoid collisions with vehicles or pedestrians. The system can also aid robots in preventing collisions while transporting goods. In tests, it outperformed traditional laser radar. Researchers aim to provide vehicles with 'X-ray vision' capabilities for high-speed travel. They plan to develop the system for various lighting conditions and automate target annotation.
Read more ›Indian Institute of Technology Develops Applications for AI and ML to Solve Engineering Problems
2019-10-29
"IIT Madras researchers have developed AISoft AI software to solve engineering problems in areas like thermal management, semiconductors, automotive, aerospace, and electronic cooling. The AI is millions of times faster than existing solutions and can handle sparse datasets independently."
Read more ›Amazon researchers reduce data required for AI transfer learning
2019-10-29
Cross-language learning is a technique where a NLP model is first trained in one language then in another. Researchers at Amazon Alexa have developed a data selection technique cutting training data in half to improve model performance. The team used multilingual embeddings and character-level embeddings along with a language model to select data for training transfer models. The transfer models outperformed those trained only on the target language, with the new data selection technique showing
Read more ›Masked Criminals: Who Else Can You Deceive?
2019-10-29
Face recognition has become a widely used technology in various sectors. New tech can identify masked individuals through 3D facial modeling. Gait recognition, not relying on faces, analyzes walking patterns to identify suspects, providing a long-distance advantage.
Read more ›Artificial Intelligence Transformation in the Automotive Industry
2019-10-29
Artificial intelligence is revolutionizing the automotive industry, enhancing in-car experiences, enabling autonomous driving, and advancing intelligent transportation. AI-driven systems improve user safety and experience with personalized settings, driver monitoring, and advanced driver-assistance features. The road to autonomous driving relies on advanced driver assistance systems and sensors like cameras, radar, and lidar. Smart cities will leverage AI to create safer transportation networks
Read more ›Facebook's New Developments in the 3D Field
2019-10-30
Advanced machine learning at Facebook can extract objects from photos in 3D. This breakthrough is crucial for AI interpretation and transforming the real world. Methods like Mesh R-CNN and C3DPO are vital for predicting 3D shapes and reconstructing models with improved accuracy and memory efficiency. This innovation will drive the field of computer vision forward to meet the demands of immersive experiences in the digital world.
Read more ›Semantic scholars at the Allen Institute can now search 175 million academic papers.
2019-10-24
Over 175 million research papers and related resources are now covered by Semantic Scholar, making it the most comprehensive academic search engine worldwide. The Allen Institute aims to establish a semantic foundation for millions of users, enabling advanced scientific search beyond keywords.
Read more ›The Road to Google's Quantum Supremacy: How Far Away is Quantum Computing from Practical Implementation?
2019-10-24
Google announced a major breakthrough in quantum computing research, achieving "quantum supremacy" by solving tasks in 200 seconds that would take a supercomputer 10,000 years. Quantum computing may revolutionize industries, but IBM's rebuttal calls for skepticism. Quantum computing has huge potential but still requires significant development and may take another 10 years to reach practical use.
Read more ›Artificial Intelligence and Healthcare: A Symbiotic Relationship
2019-10-25
"Artificial intelligence revolutionizes healthcare by providing comprehensive health information, genetic screening, and real-time monitoring of biometrics and environmental factors. AI models predict disease risks, alert emergencies, and optimize medical treatments. Intelligent robots offer smart alarm and medication adherence services. AI reduces physician burnout and expands medical services. Data integrity, legal frameworks, diverse representation, and training are essential for AI implement
Read more ›Artificial Intelligence and Information Lifecycle
2019-10-25
Record managers need to optimize RIM programs through investments in IT environments like hosting and cloud services. Implementing retention policies and AI technologies can enhance security, automate workflows, and drive organizational missions forward.
Read more ›Google Uses Deep Learning to Predict the Olfactory Properties of Molecules
2019-10-25
Smell is crucial in how organisms respond to the world; odor prediction is a multi-label classification problem. Google's GNNs significantly improve odor prediction, offering promise for future research.
Read more ›NVIDIA Aero combines GPU with 5G network to accelerate AI operations
2019-10-25
At MWC, NVIDIA CEO Jensen Huang introduced Aerial SDK for 5G RAN. Aero platform includes cuVNF for GPU-accelerated VNF and cuBB for GPU-based 5G signal processing. Red Hat's OpenShift powers NVIDIA's EGX stack for managing 5G RAN. Microsoft Azure IoT supports EGX systems for deploying AI models at scale. NVIDIA collaborates with industry leaders for end-to-end solutions in the 5G wireless market.
Read more ›Tesla to Release Beta Version of Full Self-Driving by the End of the Year
2019-10-25
Tesla CEO Elon Musk announced fully autonomous driving with pilot program launch by year end. Tesla aims for Level 4 autonomy by 2020, developing FSD chip. Musk plans to minimize driver intervention, but complete autonomy may require further optimization and regulatory approval.
Read more ›Google to Integrate Biodiversity Research AI into Tensorflow Hub
2019-10-28
Machine learning aids biodiversity research; Google partners with GBIF and Visipedia to release AI workflow for data aggregation, ensuring standardized terms and fair data coverage. Contributions to be presented at Leiden Biodiversity Conference, aiming to expand ML models globally.
Read more ›