Posts

Showing posts with the label NEURAL_NETWORKS

Can Scaling Solve Robotics?

Image
This post was originally published on the author’s personal blog. Last year’s Conference on Robot Learning (CoRL) was the biggest CoRL yet, with over 900 attendees, 11 workshops, and almost 200 accepted papers. While there were a lot of cool new ideas (see this great set of notes for an overview of technical content), one particular debate seemed to be front and center: Is training a large neural network on a very large dataset a feasible way to solve robotics?1 Of course, some version of this question has been on researchers’ minds for a few years now. However, in the aftermath of the unprecedented success of ChatGPT and other larg...... Read Full article on GretAi

Elevating Your Skills with Spot: A Focus on Reinforcement Learning from Boston Dynamics - Video

Image
Boston Dynamics has been at the forefront of legged robotics for decades, with Spot being one of their most advanced creations. In their latest video titled "Stepping Up | Reinforcement Learning with Spot | Boston Dynamics," viewers are taken behind the scenes with Paul Domanico, a Robotics Engineer on the Spot Locomotion team. In the video, Domanico discusses the team's use of reinforcement learning to enhance Spot's movement capabilities, making the robot more agile and adaptable to various environments. By combining reinforcement learning with model predictive control, Spot is able to navigate complex terrains with ease, making split-seco...... Read Full article on GretAi

9 New Examples of OpenAI's Sora - Video

Image
OpenAI's Sora model is making waves in the world of artificial intelligence, with creators and tech enthusiasts alike getting their hands on it to generate stunning videos from simple text prompts. In a recent video by tech reviewer Marquez Brownley, also known as MKBHD, we get a glimpse into the capabilities of this impressive model. The video showcases nine brand new examples of Sora in action, from a red cube being printed by a 3D printer to a product reviewer surrounded by gadgets in a well-lit environment. While Sora's generated videos still struggle with some aspects like physics and branded items, the lighting and shadows are remarkab...... Read Full article on GretAi

The Five Generations of Artificial Intelligence - Video

Image
Artificial intelligence has come a long way since its inception, evolving through five distinct generations that have shaped the landscape of AI as we know it today. The first generation, known as handmade AI, saw intelligent systems crafted by humans who relied on expert knowledge to solve problems. These systems had limited capacity for abstract thinking but were effective for specific tasks. Moving on to the second generation, statistical learning enabled computers to excel at games like chess and Go, using deep learning techniques to achieve incredible feats. The third generation focused on explainability and generative models, making AI...... Read Full article on GretAi

How effective is OpenAI's latest video model? - Video

Image
OpenAI's new video model, Sora, has been causing quite a buzz in the AI community. Many are comparing it to existing AI video tools, such as Runway, to see just how good it really is. One interesting aspect of Sora is the way it handles prompts given by users. For example, when asked to create a reflection in a window, Sora uses a new Training Method called SpaceTime patches to predict the movement of objects through time. Unlike other AI models, Sora is said to generate emergent behavior in videos without being explicitly programmed to do so. This means that it can create realistic camera movements, lighting effects, and even generate Minec...... Read Full article on GretAi

Understanding GPT4 and How to Utilize OpenAI's GPT-4 - Video

Image
GPT4 has finally arrived, leaving GPT in the dust and changing the game for artificial intelligence. This powerful AI can now process images and text, making it a multimodal platform unlike its predecessors. OpenAI recently ran a developer live stream introducing GPT4 and showcasing its capabilities, including a demo of its image to text processing. One impressive demonstration showed GPT4 accurately identifying elements in a photo and explaining the context of a joke based on a series of images. Additionally, GPT4 can convert a drawing of a website on a napkin into a functional website in just 10 to 20 seconds. This capability has wowed dev...... Read Full article on GretAi

In a spicy interview, Sam Altman reveals new details about GPT-5 - Video

Image
Sam Altman recently revealed new details about GPT-5 in a fiery interview with Axios in Davos. He expressed surprise at the overwhelming popularity of Chat GPT and GPT-4, and how they have transformed businesses and integrate into daily lives. Altman also hinted at the potential name change for GPT-5. He expressed excitement for the exponential increase in the model's intelligence and ability to solve complex problems more accurately. Regarding AI's future in 2024, Altman stressed that the emphasis will be on generalized intelligence continually increasing. He hinted at the possibility of AI being termed something other than GPT-5 and unders...... Read Full article on GretAi