Machine Learning Model
- All
- News
- Videos
-
New Physics-Based Model Sheds Light on How Deep Neural Networks Learn Features
A Physical Review Letters study likens deep neural network feature learning to spring-block mechanics, linking data simplification to spring extension and nonlinearity to friction. The model reveals how noise can balance separation across layers and help predict performance, offering a powerful tool to optimise training, improve generalisation, and...
-
YouTube’s AI Age Estimation Model to Begin Rolling Out Next Week, Will Add Restrictions to Minor Accounts
YouTube is now preparing to roll out its artificial intelligence (AI) and machine learning (ML) age estimation tool. Announced in July, this tool uses several user activity metrics to determine the real age of the user. The streaming giant says this feature will better protect minors who have provided fake dates of birth while making their accounts...
-
WeTransfer Confirms Files Not Used to Train AI Models Following Criticism, Updates Terms of Service
WeTransfer clarified it won't use user-uploaded files for AI model training, following user backlash for updates to its terms of service that suggested the platform could "reproduce, distribute, modify" content for moderation. The file transfer service has now updated its terms again, removing all references to using machine learning with user file...
-
AI Designs Ocean Gliders Inspired by Sea Creatures to Boost Underwater Research Efficiency
Researchers at MIT and the University of Wisconsin have developed AI-designed underwater gliders inspired by marine animals. The new shapes, generated using machine learning, move more efficiently underwater than traditional torpedo designs. This innovation could greatly improve climate and ocean monitoring by enabling better performance, energy sa...
-
CSIRO Uses Quantum AI to Revolutionize Semiconductor Design
In a global first, researchers at Australia’s CSIRO used quantum machine learning to enhance semiconductor design, outperforming classical AI models. By modeling Ohmic resistance in GaN transistors, the team built a hybrid quantum–classical model using just 5 qubits. This Quantum Kernel-Aligned Regressor revealed subtle fabrication patterns tha...
-
Apple Researchers Introduce Matrix3D, a Unified AI Model That Can Turn 2D Photos Into 3D Objects
Apple researchers released a new artificial intelligence (AI) model that can generate 3D views from multiple 2D images. The large language model (LLM), dubbed Matrix3D, was developed by the company’s Machine Learning team, in collaboration with Nanjing University and the Hong Kong University of Science and Technology (HKUST). The Cupertino-based ...
-
Malicious Machine Learning Models Discovered on Hugging Face: Report
Hugging Face, the artificial intelligence (AI) and machine learning (ML) hub, is said to contain malicious ML models. A cybersecurity research firm discovered two such models that contain code that can be used to package and distribute malware to those who download these files. As per the researchers, threat actors are using a hard-to-detect method...
-
OpenAI’s AI Training Opt-Out Tool for Creators Not Releasing Anytime Soon: Report
OpenAI, in May, announced a new machine learning (ML) tool that could enable creators to specify whether they wanted OpenAI to train its artificial intelligence (AI) models on their content or not. Dubbed Media Manager, the tool was said to identify copyrighted text, images, audio, and videos across multiple sources and could reflect creators’ pr...
-
Hugging Face Introduces Open-Source SmolVLM Vision Language Model Focused on Efficiency
Hugging Face, the artificial intelligence (AI) and machine learning (ML) platform, introduced a new vision-focused AI model last week. Dubbed SmolVLM (where VLM is an acronym for vision language model), it is a compact-sized model that is focused on efficiency. The company claims that due to its smaller size and high efficiency, it can be useful fo...
-
Harvard and Google Created the Most Detailed Map of Brain Using AI
In a landmark achievement, Harvard University and Google have developed the most detailed brain map ever created. This innovative project focused on a cubic millimeter of human cerebral cortex, a sample taken during an epilepsy surgery in 2014. Over a decade, biologists and machine-learning experts meticulously mapped approximately 57,000 cells and...
-
Sakana AI Introduces The AI Scientist, Claims It Can Fully Automate Scientific Discovery
Sakana AI, a Tokyo-based artificial intelligence (AI) firm, introduced its latest AI model that can conduct scientific research on Tuesday. The company claims that the AI model, dubbed The AI Scientist, can run an end-to-end scientific discovery process in the AI and machine learning (ML) fields on its own. This would include idea generation, runni...
-
Apple Claims OpenELM Does Not Power Apple Intelligence Following YouTube Video Data Controversy: Report
Apple has reportedly claimed that the OpenELM AI model it released recently was not used for its Apple Intelligence or any of the company’s machine learning features. The Cupertino-based tech giant open-sourced the OpenELM model in April. The reported statement comes just days after an investigation found that the AI model was trained on datasets...
-
EyeEm Stock Photo Marketplace Updates Its Policy to Use Uploaded Content to Train AI Models
EyeEm, the stock photo marketplace, has updated its terms and conditions to specify that the company which is presently owned by Spain-based Freepik, can use the uploaded content on the platform to train artificial intelligence (AI) and machine learning (ML) models. The platform has also told its users that in case they do not consent to this, they...
-
Google Unveils Gemini 1.5, Meta Introduces Predictive Visual Machine Learning Model V-JEPA
Google and Meta have announced new AI models. While Google has started an early testing of its Gemini 1.5 model which has the capability of long-context understanding across modalities, Meta’s Video Joint Embedding Predictive Architecture (V-JEPA) model, which has been publicly released, can teach AI models about the real world through videos and...
-
Rabbit R1 AI-Powered Pocket Assistant With 360-Degree Rotating Camera Launched: Price, Specifications
Rabbit R1 pocket AI assistant was launched by US startup Rabbit on Wednesday. It sports a 2.88-inch touchscreen display and is equipped with a scroll wheel and push-to-talk button. The device has a 360-degree rotating camera and runs on Rabbit OS that is powered by the firm's Large Action Model and can perform actions on your behalf.
-
New Physics-Based Model Sheds Light on How Deep Neural Networks Learn Features
A Physical Review Letters study likens deep neural network feature learning to spring-block mechanics, linking data simplification to spring extension and nonlinearity to friction. The model reveals how noise can balance separation across layers and help predict performance, offering a powerful tool to optimise training, improve generalisation, and...
-
YouTube’s AI Age Estimation Model to Begin Rolling Out Next Week, Will Add Restrictions to Minor Accounts
YouTube is now preparing to roll out its artificial intelligence (AI) and machine learning (ML) age estimation tool. Announced in July, this tool uses several user activity metrics to determine the real age of the user. The streaming giant says this feature will better protect minors who have provided fake dates of birth while making their accounts...
-
WeTransfer Confirms Files Not Used to Train AI Models Following Criticism, Updates Terms of Service
WeTransfer clarified it won't use user-uploaded files for AI model training, following user backlash for updates to its terms of service that suggested the platform could "reproduce, distribute, modify" content for moderation. The file transfer service has now updated its terms again, removing all references to using machine learning with user file...
-
AI Designs Ocean Gliders Inspired by Sea Creatures to Boost Underwater Research Efficiency
Researchers at MIT and the University of Wisconsin have developed AI-designed underwater gliders inspired by marine animals. The new shapes, generated using machine learning, move more efficiently underwater than traditional torpedo designs. This innovation could greatly improve climate and ocean monitoring by enabling better performance, energy sa...
-
CSIRO Uses Quantum AI to Revolutionize Semiconductor Design
In a global first, researchers at Australia’s CSIRO used quantum machine learning to enhance semiconductor design, outperforming classical AI models. By modeling Ohmic resistance in GaN transistors, the team built a hybrid quantum–classical model using just 5 qubits. This Quantum Kernel-Aligned Regressor revealed subtle fabrication patterns tha...
-
Apple Researchers Introduce Matrix3D, a Unified AI Model That Can Turn 2D Photos Into 3D Objects
Apple researchers released a new artificial intelligence (AI) model that can generate 3D views from multiple 2D images. The large language model (LLM), dubbed Matrix3D, was developed by the company’s Machine Learning team, in collaboration with Nanjing University and the Hong Kong University of Science and Technology (HKUST). The Cupertino-based ...
-
Malicious Machine Learning Models Discovered on Hugging Face: Report
Hugging Face, the artificial intelligence (AI) and machine learning (ML) hub, is said to contain malicious ML models. A cybersecurity research firm discovered two such models that contain code that can be used to package and distribute malware to those who download these files. As per the researchers, threat actors are using a hard-to-detect method...
-
OpenAI’s AI Training Opt-Out Tool for Creators Not Releasing Anytime Soon: Report
OpenAI, in May, announced a new machine learning (ML) tool that could enable creators to specify whether they wanted OpenAI to train its artificial intelligence (AI) models on their content or not. Dubbed Media Manager, the tool was said to identify copyrighted text, images, audio, and videos across multiple sources and could reflect creators’ pr...
-
Hugging Face Introduces Open-Source SmolVLM Vision Language Model Focused on Efficiency
Hugging Face, the artificial intelligence (AI) and machine learning (ML) platform, introduced a new vision-focused AI model last week. Dubbed SmolVLM (where VLM is an acronym for vision language model), it is a compact-sized model that is focused on efficiency. The company claims that due to its smaller size and high efficiency, it can be useful fo...
-
Harvard and Google Created the Most Detailed Map of Brain Using AI
In a landmark achievement, Harvard University and Google have developed the most detailed brain map ever created. This innovative project focused on a cubic millimeter of human cerebral cortex, a sample taken during an epilepsy surgery in 2014. Over a decade, biologists and machine-learning experts meticulously mapped approximately 57,000 cells and...
-
Sakana AI Introduces The AI Scientist, Claims It Can Fully Automate Scientific Discovery
Sakana AI, a Tokyo-based artificial intelligence (AI) firm, introduced its latest AI model that can conduct scientific research on Tuesday. The company claims that the AI model, dubbed The AI Scientist, can run an end-to-end scientific discovery process in the AI and machine learning (ML) fields on its own. This would include idea generation, runni...
-
Apple Claims OpenELM Does Not Power Apple Intelligence Following YouTube Video Data Controversy: Report
Apple has reportedly claimed that the OpenELM AI model it released recently was not used for its Apple Intelligence or any of the company’s machine learning features. The Cupertino-based tech giant open-sourced the OpenELM model in April. The reported statement comes just days after an investigation found that the AI model was trained on datasets...
-
EyeEm Stock Photo Marketplace Updates Its Policy to Use Uploaded Content to Train AI Models
EyeEm, the stock photo marketplace, has updated its terms and conditions to specify that the company which is presently owned by Spain-based Freepik, can use the uploaded content on the platform to train artificial intelligence (AI) and machine learning (ML) models. The platform has also told its users that in case they do not consent to this, they...
-
Google Unveils Gemini 1.5, Meta Introduces Predictive Visual Machine Learning Model V-JEPA
Google and Meta have announced new AI models. While Google has started an early testing of its Gemini 1.5 model which has the capability of long-context understanding across modalities, Meta’s Video Joint Embedding Predictive Architecture (V-JEPA) model, which has been publicly released, can teach AI models about the real world through videos and...
-
Rabbit R1 AI-Powered Pocket Assistant With 360-Degree Rotating Camera Launched: Price, Specifications
Rabbit R1 pocket AI assistant was launched by US startup Rabbit on Wednesday. It sports a 2.88-inch touchscreen display and is equipped with a scroll wheel and push-to-talk button. The device has a 360-degree rotating camera and runs on Rabbit OS that is powered by the firm's Large Action Model and can perform actions on your behalf.
