• Waymo started offering fully driverless rides to the general public in Phoenix, Arizona, the first commercial robotaxi service without a safety driver.

• Cruise received the first deployment permit for driverless taxis in California, though it has not launched the service yet.

• Tesla released a major “Full Self Driving” beta update that enabled some of its vehicles to navigate city streets with no human intervention, though it is still considered a level 2 system and requires constant driver supervision.

• Argo AI, a Ford and Volkswagen backed startup, received permits to test its autonomous vehicles on the streets of Munich, the first city outside the U.S. for the company.

• Mobileye, an Intel company, announced a robotaxi service will launch in 2022 with Safety Drivers and a fully driverless service in 2023. They plan to deploy thousands of autonomous vehicles per year in the coming years.

• Autonomous trucking continues to make progress, with companies like Plus, TuSimple, Embark and others expanding their testing and development programs.

• Sensor and lidar technology advanced, with higher resolution, further range and lower cost, helping to improve the perception capabilities of self-driving systems.

2022 saw the first commercial robotaxi services without safety drivers launch, although still in limited areas. Autonomous trucking development continues to progress. And there were advancements in the underlying technologies like sensors, lidar and AI that will feed into the next generation of self-driving systems. However, fully driverless vehicles for the mass market are still likely years away.

The post Top Technology Inventions or Innovations of 2022 for Autonomous Vehicles first appeared on Tech Like This.

• GPT-3 progress – Models like GPT-3 from OpenAI continued to improve in capabilities, now able to produce human-like text, code and even art with only a few examples or prompts.

• CLIP progress – Contrastive Language-Image Pre-Training (CLIP) models improved significantly, able to match images and text with high accuracy, enabling new types of image search and image captioning tools.

• Molformer – Google’s Molformer model achieved state-of-the-art results for molecular generation and property prediction, showing potential for drug and material discovery.

• AlphaCode – GitHub copilot’s AlphaCode model surpassed 35 billion parameters, demonstrating industry-changing capabilities for auto-completing code.

• DALL-E 2 – OpenAI released DALL-E 2, an image generator that can create high resolution images from text descriptions, achieving a major breakthrough in this field.

• Stable Diffusion – Stable Diffusion, an open source image generator comparable to DALL-E 2, was released, further democratizing access to this powerful technology.

• Gato – An AI model from Anthropic achieved human-level performance on the Arcade Learning Environment, demonstrating suprising generalization abilities from minimal training.

• Pathrust – An AI model from Google Health showed state-of-the-art performance on biomedical language tasks, indicating potential for accelerating medical research.

These represent some of the major innovations and breakthroughs in AI and machine learning in 2022, though the list is by no means exhaustive. As you can see, progress was made across a wide range of applications, from text generation to image generation to molecular and code generation. The capabilities and potential of these technologies continue to grow at an astounding pace.

The post The Top Technology Inventions or Innovations of 2022 for Artificial Intelligence and Machine Learning first appeared on Tech Like This.