🚀 Medial Secures Investment on Shark Tank India - Fueling the Future of Professional Social Networking. 🔥

Startup Showcase

Premium Content

Try our Valuation Calculator →

News on Medial

Large language models can help home robots recover from errors without human help

TechCrunch · 1y ago

Large language models can help home robots recover from errors without human help

MIT researchers have developed a method to help robots automatically correct mistakes without requiring human intervention. The researchers used large language models (LLMs) to eliminate the need for programmers to label and assign subactions manually. By breaking demonstrations into smaller subsets, the LLMs can identify and label each step of a task, allowing the robot to replan and recover on its own when faced with environmental variations or small disturbances. The system proved successful in a demonstration where a robot was trained to scoop marbles and pour them into a bowl, self-correcting mistakes along the way.

Related News

The man who made robots dance now wants them to think for themselves

Wired · 1y ago

The man who made robots dance now wants them to think for themselves

The Boston Dynamics AI Institute, founded by Marc Raibert, is focused on developing robots that can comprehend and handle complex situations without human assistance. While Boston Dynamics' robots have garnered attention for their physical capabilities, the institute aims to enhance their cognitive abilities. Raibert discusses the institute's mission and its efforts to incorporate language models and reinforcement learning into robotics. The goal is to advance AI beyond large language models and explore the integration of physical and cognitive intelligence. The institute is also working on projects such as bicycle repair robots and dexterous manipulation. Raibert addresses concerns about robots and AI becoming dangerous, emphasizing that most bad things in the world do not come from robots.

OpenAI’s new “CriticGPT” model is trained to criticize GPT-4 outputs

Arstechnica

Arstechnica · 1y ago

OpenAI’s new “CriticGPT” model is trained to criticize GPT-4 outputs

OpenAI has introduced CriticGPT, an AI model designed to detect errors in code generated by ChatGPT. The model aims to improve the alignment of AI systems with human expectations by employing Reinforcement Learning from Human Feedback (RLHF). CriticGPT, based on the GPT-4 family of large language models (LLMs), assists human reviewers by identifying coding mistakes in the output. The researchers trained CriticGPT on a dataset of intentionally bugged code to teach it how to recognize and flag different errors. According to experiments, CriticGPT's critiques were preferred by trainers over human critiques in 63% of cases involving LLM errors.

Toyota's robots are learning to do housework—By copying humans

Wired · 1y ago

Toyota's robots are learning to do housework—By copying humans

Toyota Research Institute is developing robots that can learn household chores by combining machine learning techniques and language models. The robots are trained using a diffusion policy machine learning system, similar to ones used in AI image generators, to determine their actions based on various data sources. Toyota aims to utilize language models to enable robots to learn tasks by watching videos, potentially from platforms like YouTube. The goal is to have robots adapt, improvise, and flexibly handle tasks that require human-like abilities.

WTF is AI?

TechCrunch · 1y ago

WTF is AI?

Artificial intelligence (AI) is software that mimics human thinking and can be useful in getting tasks done. It approximates human thinking by detecting and continuing patterns. AI models, like large language models, map out language patterns and can generate responses based on those patterns. AI models are good at low-value written work, coding tasks, sorting and summarizing information, and even engaging in conversations. However, AI models have limitations such as generating hallucinations, bias, and the need for proper training data. The concept of artificial general intelligence (AGI), though not currently achievable, raises questions about potential benefits and risks.

Large AI models can now create smaller AI tools without humans and train them like a 'big brother,' scientists say

Business Insider

Business Insider · 1y ago

Large AI models can now create smaller AI tools without humans and train them like a 'big brother,' scientists say

Researchers from MIT, University of California campuses, and AI technology company Aizip claim that large AI models can now autonomously create smaller AI tools without human assistance. These smaller models, which can be generated by more advanced AI models like the one used by ChatGPT, have the potential to greatly impact our daily lives. The team views this as the first step towards self-evolving AI and believes that in the future, large and small AI models will collaborate to build a complete intelligence ecosystem.

Three reasons why it’s good news that robots are getting smarter

Livemint

Livemint · 1y ago

Three reasons why it’s good news that robots are getting smarter

Advances in robotics, driven by artificial intelligence (AI), are revolutionizing the field. Robots are becoming more capable, easier to program, and better at explaining their actions. Multimodal AI models combine language and vision understanding with robotic sensor data, enabling robots to understand and respond to human commands. Additionally, these models allow robots to provide reasoning behind their actions, making programming and debugging more straightforward. Improved learning capabilities and generalization from one skill to another are also enabling robots to move beyond factories and warehouses to assist with tasks like cooking, cleaning, and caring for the elderly.

Can't write code without LLMs anymore: Zerodha CTO Kailash Nadh

Inshorts · 8m ago

Can't write code without LLMs anymore: Zerodha CTO Kailash Nadh

Zerodha CTO Kailash Nadh said he can't write code without the help of large language models anymore. "LLMs have become like a second brain for developers," he told Moneycontrol. "I'm even learning a new language in a way that was not possible before," he said, adding, "We (Zerodha) use it heavily for code generation, not massive amounts, but for...focused tasks."

British officials say AI chatbots could carry cyber risks

Economic Times

Economic Times · 1y ago

British officials say AI chatbots could carry cyber risks

In a pair of blog posts due to be published Wednesday, Britain's National Cyber Security Centre (NCSC) said that experts had not yet got to grips with the potential security problems tied to algorithms that can generate human-sounding interactions - dubbed large language models, or LLMs.

OpenAI introduces CriticGPT, a GPT 4-based model that can detect errors in ChatGPT’s code output

Money Control

Money Control · 1y ago

OpenAI introduces CriticGPT, a GPT 4-based model that can detect errors in ChatGPT’s code output

OpenAI has launched CriticGPT, an AI tool designed to identify errors made by the popular chatbot, ChatGPT. Unlike other language models, CriticGPT is specifically built to provide feedback and critique on ChatGPT responses. It aims to help human trainers and coders spot mistakes during the reinforcement learning process. OpenAI claims that CriticGPT improves code review outcomes by over 60% compared to previous models. However, CriticGPT currently has limitations in generating longer critiques and is trained on shorter ChatGPT answers. OpenAI has plans to expand its capabilities and address these limitations in the future.

Former Salesforce exec says large language models must learn how to code if they want to level up

Business Insider

Business Insider · 1y ago

Former Salesforce exec says large language models must learn how to code if they want to level up

Richard Socher, former chief scientist at Salesforce and CEO of You.com, believes that large language models (LLMs) can be improved by teaching them to respond to prompts in code, instead of simply predicting the next token. While LLMs exhibit impressive skills, they can still produce factual errors when asked complex mathematical questions. Socher suggests translating questions into computer code to generate accurate answers. He argues that programming will provide these models with more capabilities in the future, as scaling them with more data might not yield significant improvements.

Trackers

Active Indian VC’s

OG Capital Email

With a hands-on approach, OG Capital aims to invest in over 20 promising...

Accel Partners Email

Early and growth-stage investments in disruptive technology companies with...

Early-stage venture capital firm investing in technology startups in India. Focus on...

Access All Trackers

Startup Showcase Winners

June 2025

Helping your parents when you are miles away

The Pit Stop Your Cravings Deserve

The next generation E-commerce platform

Enter Ongoing Startup Showcase

Top Users

Trending News on Medial

30 Startups To Watch: Startups Th ...

GitHub's CEO Drops a Bombshell: " ...

Validation, loneliness, insecurit ...

Download the medial app to read full posts, comements and news.