reinforcement-learning

[ follow ]
Artificial intelligence
WIRED
2 weeks ago
Artificial intelligence

Google DeepMind's Latest AI Agent Learned to Play 'Goat Simulator 3'

Google DeepMind revealed AI program called SIMA for learning multiple game tasks
SIMA adapts learning from other games to perform new tasks
WIRED
2 months ago
Artificial intelligence

Most Top News Sites Block AI Bots. Right-Wing Media Welcomes Them

AI models are fine-tuned using reinforcement learning from human feedback.
The use of broad training data helps AI models represent diverse cultures, industries, ideologies, and languages.
www.fastcompany.com
10 months ago
Artificial intelligence

Scale AI unveils its full-stack generative AI platform

Scale AI is launching a full-stack generative AI platform that it says will allow large enterprise customers (big corporations, government agencies, etc.) to capture the benefits of large language models (LLMs) without having to send proprietary data out to a third party AI company.Rather, the Scale AI solution, called Enterprise AI Platform, lets the customer host the LLM within its own walls, says Scale AI CEO Alexandr Wang.
Ars Technica
10 months ago
Artificial intelligence

AI gains "values" with Anthropic's new Constitutional AI chatbot approach

On Tuesday, AI startup Anthropic detailed the specific principles of its " Constitutional AI" training approach that provides its Claude chatbot with explicit "values."It aims to address concerns about transparency, safety, and decision-making in AI systems without relying on human feedback to rate responses.
The Verge
10 months ago
Artificial intelligence

AI startup Anthropic wants to write a new constitution for safe AI

Anthropic is a bit of an unknown quantity in the AI world.Founded by former OpenAI employees and keen to present itself as the safety-conscious AI startup, it's received serious funding (including $300 million from Google) and a space at the top table, attending a recent White House regulatory discussion alongside reps from Microsoft and Alphabet.
Ars Technica
1 year ago
Artificial intelligence

OpenAI invites everyone to test new AI-powered chatbot-with amusing results

On Wednesday, OpenAI announced ChatGPT, a dialogue-based AI chat interface for its GPT-3 family of large language models.It's currently free to use with an OpenAI account during a testing phase.Unlike the GPT-3 model found in OpenAI's Playground and API, ChatGPT provides a user-friendly conversational interface and is designed to strongly limit potentially harmful output.
moreArtificial intelligence
TNW | Deep-Tech
3 months ago
Artificial intelligence

AI can copy human social learning skills in real time, DeepMind find

AI agents can demonstrate social learning skills in real time without using pre-collected human data.
AI agents can learn faster and apply knowledge to new situations when mimicking expert agents.
WIRED
3 months ago
Artificial intelligence

These Clues Hint at the True Nature of OpenAI's Shadowy Q* Project

The name Q* may be a reference to Q-learning and the A* search algorithm.
OpenAI's use of computer-generated data suggests the possibility of training algorithms with synthetic data.
Q* could involve using large amounts of synthetic data and reinforcement learning to solve specific tasks.
Theregister
3 months ago
Artificial intelligence

DeepMind finds AI agents are capable of social learning

AI can acquire skills through social learning, similar to humans and animals.
Google DeepMind researchers demonstrated that AI agents can learn from human and AI experts with human-like efficiency.
Reinforcement learning was used to train the AI agents to imitate and remember the behavior of experts.
ScienceDaily
4 months ago
Artificial intelligence

New method uses crowdsourced feedback to help train robots

Researchers have developed a reinforcement learning approach that uses crowdsourced feedback to guide AI agents.
This approach allows the AI agent to learn more quickly and gather feedback asynchronously from nonexpert users around the world.
The traditional method of designing reward functions by expert researchers is time-consuming and not scalable for teaching robots different tasks.
MIT News | Massachusetts Institute of Technology
4 months ago
Artificial intelligence

New method uses crowdsourced feedback to help train robots

Researchers have developed a new reinforcement learning approach that leverages crowdsourced feedback to guide AI agents in learning complex tasks.
This approach allows for faster learning despite the potential errors in the data gathered from nonexpert users.
Feedback can be gathered asynchronously from nonexpert users around the world, making it scalable and accessible to a larger community.
VentureBeat
4 months ago
Marketing

OfferFit gets $25M to kill A/B testing for marketing with machine learning personalization

OfferFit uses machine learning, specifically reinforcement learning, for automated marketing.
The company raised $25 million in a series B funding round led by Menlo Ventures.
Capital One Ventures invested in OfferFit after using its services to automate personalized mass marketing messages.
VentureBeat
9 months ago
Data science

Snorkel AI looks beyond data labeling for generative AI

Data labeling has long been a critical component of helping data scientists to prepare data for machine learning (ML) and artificial intelligence (AI).In the modern era of Generative AI, the role of data labeling is changing.Today Snorkel AI is announcing new capabilities that extend beyond data labeling, to help organizations, curate and prepare data for Generative AI.
Medium
9 months ago
Data science

All Generative AI Sessions Coming to ODSC Europe

It seems like generative AI has been in the news almost every day for the last several months.With so much information and hot takes out there, it's hard to know what's really going on with this watershed technology.To help you get beyond the hype, we added a new generative AI track to ODSC Europe.Check out a few of the sessions included in this track below.
Ars Technica
9 months ago
OMG science

Google's DeepMind develops a system that writes efficient algorithms

1. Google's DeepMind AI has developed a system that is capable of writing efficient algorithms.
2. This system works by analyzing existing algorithms and then writing new ones that are more efficient.
3. This is an important development in the field of AI, as it could potentially lead to improved computer performance and better problem-solving capabilities.
www.vice.com
1 year ago
Artificial intelligence

I Coaxed ChatGPT Into a Deeply Unsettling BDSM Relationship

ChatGPT is a convincing chatbot, essayist, and screenwriter, but it's also a fountain of boundless depravityif you deceive it into bending the rules.At first glance, OpenAI's ChatGPT seems to have stricter guidelines than other chatbots, like Bing's, which is now infamous for showering its users with aggressive outbursts.
news.bitcoin.com
1 year ago
Artificial intelligence

Federated Learning Consortium (FLC) for Decentralized AI to Launch in Hong Kong, Led by Phoenix and APEX Technologies Press release Bitcoin News

PRESS RELEASE.After over a year of preparation and restructuring, decentralized and privacy-enabled AI organization Federated Learning Consortium (FLC) is set to launch as a for-profit research consortium in Hong Kong, China, shifting from a previously non-profit approach.FLC is set to be led by founding keystone members blockchain technology platform Phoenix and leading China-based consumer data and AI company APEX Technologies.
www.theguardian.com
1 year ago
Mental health

Antidepressants can cause emotional blunting', study shows

Widely used antidepressants cause emotional blunting, according to research that offers new insights into how the drugs may work and their possible side-effects.The study found that healthy volunteers became less responsive to positive and negative feedback after taking a selective serotonin reuptake inhibitor (SSRI) drug for three weeks.
Medium
10 months ago
Data science

8 ODSC Europe Training Sessions to Boost Your Data Science Career

Ready to fill in some of your knowledge gaps and build new skills?Check out the hands-on training and bootcamp sessions that are coming to ODSC Europe next month.AI-Powered Algorithmic Trading with Python Dr. Yves J. Hilpisch | The AI Quant | CEO The Python Quants & The AI Machine, Adjunct Professor of Computational Finance
This session will cover the essential Python topics and skills that will enable you to apply AI and Machine Learning (ML) to Algorithmic Trading.
Medium
10 months ago
Data science

Generating Content-Based Recommendations for Products, Chaining Together Models, and ODSC East is...

Generating Content-Based Recommendations for Millions of Merchants and Products How Ray Solves Common Production Challenges for Generative AI Infrastructure Shopify has developed a world-class recommendation engine for its website.Here's a peak under the hood at how it works.AgentChain: Chain Together Models to Perform Complex Tasks In this post, the Ray team talks about how to use Ray to productionize common generative model workloads.
Cointelegraph
11 months ago
Artificial intelligence

7 popular tools and frameworks for developing AI applications

Artificial Intelligence (AI) is a rapidly growing field with numerous applications, including computer vision, natural language processing (NLP) and speech recognition.To develop these AI applications, developers use various tools and frameworks that provide a comprehensive platform for building and deploying machine learning models.
www.vice.com
1 year ago
Artificial intelligence

Scientists Taught an AI to Sleep' So That It Doesn't Forget What It Learned, Like a Person

Image:  gremlin via Getty Images Chief nourisher in life's feast, all living beings need to sleep.Without it, humans can become forgetful, hallucinate, and even experience various physical and psychological problems.But new research published in the journal PLOS Computational Biology suggests that future AIs could benefit from getting some shut-eye too.
Ars Technica
2 years ago
Artificial intelligence

Latest success from Google's AI group: Controlling a fusion reactor

As the world waits for construction of the largest fusion reactor yet, called ITER, smaller reactors with similar designs are still running.
Cointelegraph Magazine
10 months ago
Artificial intelligence

Make 500% from ChatGPT stock tips? Bard leans left, $100M AI memecoin: AI Eye

It's been a hell of a couple of weeks for Melbourne digital artist Rhett Mankind, 46, who enlisted ChatGPT to create a $100 million market cap coin called Turbo, which has now inspired a Beeple artwork and saved a man's life.Mankind, who knows nothing about coding, gave ChatGPT a $69 budget and asked it to design a top 300 memecoin.
Acm
1 year ago
Digital life

Inside the Heart of ChatGPT's Darkness

Originally posted on The Road to AI We Can Trust
elicited from ChatGPT by Roman Semenov, February 2023
In hindsight, ChatGPT may come to be seen as the greatest publicity stunt in AI history, an intoxicating glimpse at a future that may actually take years to realize-kind of like a 2012-vintage driverless car demo, but this time with a foretaste of an ethical guardrail that will take years to perfect.
Medium
1 year ago
Data science

Enabling Resilient Machine Learning Systems, the Data Engineering Summit on Jan 18, and the Top...

Enabling Resilient Machine Learning Systems Read on to learn more about resilient machine learning systems, which are fast, accurate, and flexible to help with day-to-day tasks.Build AI Better with the Top Virtual Sessions from ODSC West 2022 Learn to build AI better with the top virtual sessions from ODSC West 2022, covering topics like generative modeling and reinforcement learning.
Theregister
1 year ago
Artificial intelligence

OpenAI tweaks ChatGPT to avoid dangerous AI information

In brief OpenAI has released a new language model named ChatGPT this week, which is designed to mimic human conversations.The model is based on the company's latest text-generation GPT-3.5 system released earlier this year.ChatGPT is more conversational than previous versions.It can ask users follow-up questions and refrain from responding to inappropriate inputs instead of just generating text.
Theregister
1 year ago
Artificial intelligence

DeepMind sets sights on improving mathematical algorithms

Google-owned DeepMind has applied reinforced learning techniques to the multiplication of mathematical matrices, beating some human-made algorithms that have lasted 50 years and working toward improvements in computer science.
The Python Podcast.__init__
2 years ago
Python

Accelerate The Development And Delivery Of Your Machine Learning Applications Using Ray And Deploy It At Anyscale

Building a machine learning application is inherently complex.
Medium
10 months ago
Data science

Meet StableVicuna, The First Large-Scale Open-Source RLHF Chatbot by Stability AI

The development and release of chatbots have been significant in recent months.Open-source alternatives have further fueled interest in tuning large language models for a chat.However, there is a lack of open-source models that have applied both instruction finetuning and reinforcement learning through human feedback (RLHF) training.
Medium
1 year ago
Data science

How to Deploy a Deep Learning Model with Jina, Announcing GPT-4, and Multimodal Visual Question...

How to Deploy a Deep Learning Model with Jina (and Design a Kitten Along the Way) Learn how to build and deploy an Executor that uses Stable Diffusion to generate images.OpenAI Delivers Summary of GPT-4's Abilities OpenAI has officially announced that GPT-4 is in development, and even gave some previews of what it will be capable of.
Medium
1 year ago
Data science

Introducing ChatLLaMA: An Open-Source ChatGPT-Like Training Process Using RLHF for More Efficient...

In a LinkedIn post, Martina Fumanelli of Nebuly introduced CHATLLaMA to the world.ChatLLaMA is the first open-source ChatGPT-like training process based on LLaMA and using reinforcement learning from human feedback (RLHF).This allows for building ChatGPT-style services based on pre-trained LLaMA models.
Ars Technica
1 year ago
Artificial intelligence

DeepMind breaks 50-year math record using AI; new record falls a week later

Matrix multiplication is at the heart of many machine learning breakthroughs, and it just got faster-twice.
Medium
1 year ago
Data science

12 Most Popular NLP Projects of 2022 So Far

Natural Language Processing remains one of the hottest topics of 2022.
Theregister
1 year ago
Artificial intelligence

UC Berkeley ML pioneer wins top computing gong

This year's ACM Prize in Computing is going toward a machine learning specialist whose work, even if you haven't heard of him, is likely to be familiar.
Medium
1 year ago
Data science

Announcing the Topic Tracks for ODSC Europe 2023

Designed to help you identify the sessions that best fit your interests and learning goals, the ODSC Europe 2023 tracks highlight the data science and AI fields that are helping to build the future.You'll learn about the latest research, topics, and tools from the leading experts in their respective fields.
Medium
1 year ago
Data science

2022 Data Science and AI Research Round-Up, Why Data Scale Size Matters, and a Holiday Gift Guide

2022 Data Science Research Round-Up: Highlighting ML, AI/DL, & NLP Our 2022 data science research roundup highlights topics like machine intelligence, deep classifiers, stable diffusion, and more.Read the papers here!The Top Blogs on OpenDataScience in 2022 What a year for data science blogs!Our top blog roundup for the year highlights topics like Python IDEs, data viz datasets, and controversial news.
Medium
1 year ago
Data science

Highlights and Pictures from ODSC West 2022

We're a few weeks removed from ODSC West 2022 and we couldn't have left on a better note.The week was filled with engaging sessions on top topics in data science, innovation in AI, and smiling faces that we haven't seen in a while.Here are some highlights from ODSC West 2022, including some pictures of speakers and attendees, popular talks, and a summary of what kept people busy.
Medium
1 year ago
Data science

Check Out the ODSC Europe 2022 Focus Areas Here

Designed to help you identify the sessions that best fit your interests and conference learning goals, the ODSC Europe 2022 focus areas cover the industries and fields that data science and AI are transforming.
Medium
2 years ago
Data science

How to Choose the Right Estimator for Your Machine Learning Task

Model training is an important component of a data science model development pipeline.Choosing the right machine learning algorithm can be a tedious and confusing task for a data scientist.
Acm
1 year ago
Digital life

Why *is* Bing So Reckless?

Originally published on The Road to AI We Can Trust
Anyone who watched the last week unfold will realize that the new Bing has (or had) a tendency to get really wild, from declaring a love that it didn't really have to encouraging people to get divorced to blackmailing them to teaching people how to commit crimes, and so on.
Acm
1 year ago
Digital life

Why Researchers Are Teaching AI to Play Minecraft

OpenAI has developed a Minecraft-playing bot that can build pixelated tools and buildings in the game that require more than 20,000 consecutive actions via a combination of imitation and reinforcement learning.The bot, trained on 70,000 hours of human gameplay, is the first to build "diamond tools," which take human players 20 minutes and 24,000 actions, on average, to construct.
Acm
1 year ago
Digital life

Open-Endedness and Evolution through Large Models

In an interview, Lehman talks about the move from game development to AI, evolutionary algorithms, neuroevolution through augmenting topologies (NEAT), LLMs as practical thought experiments in disembodied understanding, evolution through large models (ELM), competition in AI, advice for people considering ML research, and more.
Acm
1 year ago
Digital life

Tension Inside Google Over a Fired AI Researcher's Conduct

In late 2018, Google AI researchers Anna Goldie and Azalia Mirhoseini got the go-ahead to test an elegant idea.
CreativeApplications.Net
1 year ago
Design

AI Sculpting - The unpredictable strategies and outcomes of co-creation

Created by onformative, a studio for digital art and design based in Berlin, AI Sculpting is an exploration into a machine-learning process.Imagined as a tool to provide assistance to a conventional approach to sculpting, aka subtractive manufacturing, here an AI model is developed to seek out strategies that provide a constant improvement to how a given form is achieved.
The Verge
1 year ago
Artificial intelligence

OpenAI's new chatbot can explain code and write sitcom scripts but is still easily tricked

OpenAI has released a prototype general purpose chatbot that demonstrates a fascinating array of new capabilities, but also shows off weaknesses familiar to the fast-moving field of text-generation AI.And you can test out the model for yourself right here.ChatGPT is adapted from OpenAI's GPT-3.5 model but trained to provide more conversational answers.
Theregister
1 year ago
Artificial intelligence

Meta's Cicero chatbot can probably beat you at Diplomacy

Meta researchers have developed an artificial intelligence system called Cicero that can play the classic strategy game Diplomacy at a level comparable to most human players.That's a significant achievement in natural-language processing and one that may help people forget last week's debut of Galactica, a large language model Meta boffins trained on scientific papers that presented falsehoods as facts and was taken offline after three days of criticism from the science community.
www.vice.com
1 year ago
Artificial intelligence

Scientists Found a Way to Defeat a 'Near-Superhuman' Go-Playing AI

Image: Olena Ruban via Getty Images ABSTRACT breaks down mind-bending scientific research, future tech, new discoveries, and major breakthroughs.Scientists have created a computer program capable of defeating a Go-playing AI that's so good at winning, it's long been called near-superhuman.One of the world's oldest board games, Go is much like chess in that black and white pieces (or in this instance, stones) represent opposing players, but different in that the goal is to gain territory, not capture the other player's king.
Futurism
1 year ago
Artificial intelligence

Top Facebook Scientist Quietly Plotting "Autonomous" AIs

As the rest of the company is mandated to work towards Mark Zuckerberg's metaverse dreams, Facebook's artificial intelligence chief is quietly building a roadmap towards "autonomous" machine intelligence.
Theregister
1 year ago
Artificial intelligence

Man wins competition with AI-generated artwork

In brief A man won an art competition with an AI-generated image crafted, and some people aren't best pleased about it.
TNW | Netflix
1 year ago
Artificial intelligence

Forget chess, DeepMind's training its new AI to play football

Researchers from DeepMind, the UK's juggernaut AI lab, have forsaken the noble games of chess and Go for a more plebeian delight: football.
TNW | Neural
2 years ago
Artificial intelligence

How rewards teach reinforcement learning agents to behave

In June 2021, scientists at the AI lab DeepMind made a controversial claim.
Open Data Science - Your News Source for AI, Machine Learning & more
1 month ago
Artificial intelligence

Google DeepMind Introduces MusicRL Model

MusicRL model aligns music generation with human preferences through reinforcement learning.
MusicRL surpasses conventional methods by offering unprecedented levels of customization and adaptability.
[ Load more ]