The AI Community On Kbin
- Startup Stability AI CEO Emad Mostaque Steps Downwww.bloomberg.com Startup Stability AI CEO Emad Mostaque Steps Down
Stability AI Chief Executive Officer Emad Mostaque has resigned from the British artificial intelligence startup — a move that follows quarrels with investors and waves of senior staff departures.
Stability AI Chief Executive Officer Emad Mostaque has resigned from the British artificial intelligence startup — a move that follows quarrels with investors and waves of senior staff departures.
- Chat Evolution: Top ChatGPT Alternatives for 2024trendvale.com Chat Evolution: Top ChatGPT Alternatives for 2024
Numerous ChatGPT alternatives offer superior solutions for your business, addressing diverse needs such as marketing, sales, ideation, revision.
Numerous ChatGPT alternatives offer superior solutions for your business, addressing diverse needs such as marketing, sales, ideation, revision.
- AI in Manufacturing: Applications, Examples & Benefitswww.ediiie.com AI in Manufacturing: Applications, Examples & Benefits
AI in manufacturing sees the convergence of smart computing and algorithms with tasks that dictate the intricacies of the work.
AI in manufacturing sees the convergence of smart computing and algorithms with tasks that dictate the intricacies of the work.
- AI in Telemedicine: How AI is Revolutionizing Patient Carewww.ediiie.com AI in Telemedicine: How AI is Revolutionizing Patient Care
While telemedicine doesn’t encompass every facet of AI, the role of AI in telemedicine has notably expanded in the recent past. Let's understand its use cases and benefits.
While telemedicine doesn’t encompass every facet of AI, the role of AI in telemedicine has notably expanded in the recent past. Let's understand its use cases and benefits.
- Hands-on with Gemini: Interacting with multimodal AI
YouTube Video
Click to view this content.
Gemini is our natively multimodal AI model capable of reasoning across text, images, audio, video and code. This video highlights some of our favorite intera...
- Why it’ll be hard to tell if AI ever becomes consciouswww.technologyreview.com Why it’ll be hard to tell if AI ever becomes conscious
We know remarkably little about how AI systems work, so how will we know if AI becomes conscious?
We know remarkably little about how AI systems work, so how will we know if AI becomes conscious?
- Generative AI: Use Cases, Applications and Toolswww.ediiie.com Generative AI: Use Cases, Applications and Tools
Under generative AI, tools and programs are coded to use AI in order to craft new bits and types of content. Let's explore the benefits, uses cases, and application of generative AI in various industries.
Under generative AI, tools and programs are coded to use AI in order to craft new bits and types of content. Let's explore the benefits, uses cases, and application of generative AI in various industries.
- Training AI to play Pokemon with reinforcement learning
YouTube Video
Click to view this content.
- Artificial Intelligence Could Finally Let Us Talk with Animalswww.scientificamerican.com Artificial Intelligence Could Finally Let Us Talk with Animals
AI is poised to revolutionize our understanding of animal communication
AI is poised to revolutionize our understanding of animal communication
- Researchers discover 'Reversal Curse:' LLMs trained on "A is B" fail to learn "B is A"
Training AI models like GPT-3 on "A is B" statements fails to let them deduce "B is A" without further training, exhibiting a flaw in generalization. (https://arxiv.org/pdf/2309.12288v1.pdf)
Ongoing Scaling Trends
-
10 years of remarkable increases in model scale and performance.
-
Expects next few years will make today's AI "pale in comparison."
-
Follows known patterns, not theoretical limits.
No Foreseeable Limits
-
Skeptical of claims certain tasks are beyond large language models.
-
Fine-tuning and training adjustments can unlock new capabilities.
-
At least 3-4 more years of exponential growth expected.
Long-Term Uncertainty
-
Can't precisely predict post-4-year trajectory.
-
But no evidence yet of diminishing returns limiting progress.
-
Rapid innovation makes it hard to forecast.
TL;DR: Anthropic's CEO sees no impediments to AI systems continuing to rapidly scale up for at least the next several years, predicting ongoing exponential advances.
-
- RAIN: Your Language Models Can Align Themselves without Finetuning - Microsoft Research 2023 - Reduces the adversarial prompt attack success rate from 94% to 19%! AI
Paper: https://arxiv.org/abs/2309.07124
Abstract:
> > > Large language models (LLMs) often demonstrate inconsistencies with human preferences. Previous research gathered human preference data and then aligned the pre-trained models using reinforcement learning or instruction tuning, the so-called finetuning step. In contrast, aligning frozen LLMs without any extra data is more appealing. This work explores the potential of the latter setting. We discover that by integrating self-evaluation and rewind mechanisms, unaligned LLMs can directly produce responses consistent with human preferences via self-boosting. We introduce a novel inference method, Rewindable Auto-regressive INference (RAIN), that allows pre-trained LLMs to evaluate their own generation and use the evaluation results to guide backward rewind and forward generation for AI safety. Notably, RAIN operates without the need of extra data for model alignment and abstains from any training, gradient computation, or parameter updates; during the self-evaluation phase, the model receives guidance on which human preference to align with through a fixed-template prompt, eliminating the need to modify the initial prompt. Experimental results evaluated by GPT-4 and humans demonstrate the effectiveness of RAIN: on the HH dataset, RAIN improves the harmlessness rate of LLaMA 30B over vanilla inference from 82% to 97%, while maintaining the helpfulness rate. Under the leading adversarial attack llm-attacks on Vicuna 33B, RAIN establishes a new defense baseline by reducing the attack success rate from 94% to 19%. > >
- [Research Paper] Meta: Chain-of-Verification Reduces Hallucination in Large Language Models
https://arxiv.org/abs/2309.11495
Abstract
Generation of plausible yet incorrect factual information, termed hallucination, is an unsolved issue in large language models. We study the ability of language models to deliberate on the responses they give in order to correct their mistakes. We develop the Chain-of-Verification (CoVe) method whereby the model first (i) drafts an initial response; then (ii) plans verification questions to fact-check its draft; (iii) answers those questions independently so the answers are not biased by other responses; and (iv) generates its final verified response. In experiments, we show CoVe decreases hallucinations across a variety of tasks, from list-based questions from Wikidata, closed book MultiSpanQA and longform text
https://i.imgur.com/TDXcdMI.jpeg
https://i.imgur.com/XfRVxJT.jpeg
Conclusion
We introduced Chain-of-Verification (CoVe), an approach to reduce hallucinations in a large language model by deliberating on its own responses and self-correcting them. In particular, we showed that models are able to answer verification questions with higher accuracy than when answering the original query by breaking down the verification into a set of simpler questions. Secondly, when answering the set of verification questions, we showed that controlling the attention of the model so that it cannot attend to its previous answers (factored CoVe) helps alleviate copying the same hallucinations. Overall, our method provides substantial performance gains over the original language model response just by asking the same model to deliberate on (verify) its answer. An obvious extension to our work is to equip CoVe with tool-use, e.g., to use retrieval augmentation in the verification execution step which would likely bring further gains.
- 'This is the last opportunity for us to wake up': A leading economist warns we're headed for an AI-driven cataclysmwww.businessinsider.com 'This is the last opportunity for us to wake up': A leading economist warns we're headed for an AI-driven cataclysm
How Daron Acemoglu, one of the world's most respected experts on the economic effects of technology, learned to start worrying and fear AI.
How Daron Acemoglu, one of the world's most respected experts on the economic effects of technology, learned to start worrying and fear AI.
- Indeed's CEO says college students might be learning skills that could go 'obsolete' once they graduate — all because of AIwww.businessinsider.com Indeed's CEO says college students might be learning skills that could go 'obsolete' once they graduate — all because of AI
Will everything you learned in college be replaced by ChatGPT? The CEO of job site Indeed says it's not out of the question.
Will everything you learned in college be replaced by ChatGPT? The CEO of job site Indeed says it's not out of the question.
- I Asked AI to Generate an Animated Liminal Spaceship
YouTube Video
Click to view this content.
Subscribe for good luck - Most Popular videos - https://www.youtube.com/watch?v=72qmtr41iWU&list=UULP3MF3KCtKQkCudSvNpk3BOg#liminalspace , #ai , #aiart \_Arti...
- Retro Psychedelic Mummy Generated by AI
YouTube Video
Click to view this content.
Subscribe for good luck - Most Popular videos - https://www.youtube.com/watch?v=72qmtr41iWU&list=UULP3MF3KCtKQkCudSvNpk3BOg#aimusic , #ai , #aiart \_Artificia...
- Google’s Bard AI can now access Gmail, Drive, Docs, and morewww.pcworld.com Google’s Bard AI can now access Gmail, Drive, Docs, and more
The latest updates to Google’s generative AI chat bot lets it dig through your personal email, documents, and more—so you can get things done faster.
The latest updates to Google’s generative AI chat bot lets it dig through your personal email, documents, and more—so you can get things done faster.
- Nearly half of CEOs believe AI could replace their own jobs, says new poll—and 47% say that's a good thingwww.cnbc.com Nearly half of CEOs believe AI could replace their own jobs, says new poll—and 47% say that's a good thing
Many American CEOs say they're worried about their workplace's lack of AI skills, a new survey of C-suite executives and workers found. Here's why.
Many American CEOs say they're worried about their workplace's lack of AI skills, a new survey of C-suite executives and workers found. Here's why.
- China aims to replicate human brain in bid to dominate global AIwww.newsweek.com China aims to replicate human brain in bid to dominate global AI
The pursuit of the most advanced AI—human-like artificial general intelligence—has prompted concerns among experts about potential dangers if it runs amok.
The pursuit of the most advanced AI—human-like artificial general intelligence—has prompted concerns among experts about potential dangers if it runs amok.
- AI that’s smarter than humans? Americans say a firm “no thank you.”
63 percent of Americans want regulation to prevent artificial general intelligence, or AGI, which OpenAI aims to build.
- Trippy Video Generated by AI (Spectrophobic)
YouTube Video
Click to view this content.
Subscribe for good luck - Most Popular videos - https://www.youtube.com/watch?v=72qmtr41iWU&list=UULP3MF3KCtKQkCudSvNpk3BOg#chatgpt , #ai , #art\_Artificial I...
- 9 free AI tools that run locally on your PCwww.pcworld.com 9 free AI tools that run locally on your PC
These clever AI tools can have a big impact by using elaborate models to tackle demanding tasks. The nine programs presented here have something in common besides AI: they are freely available.
These clever AI tools can have a big impact by using elaborate models to tackle demanding tasks. The nine programs presented here have something in common besides AI: they are freely available.
- Fewer workers, higher pay, robots: aerospace suppliers push automationwww.seattletimes.com Fewer workers, higher pay, robots: aerospace suppliers push automation
Two Seattle-area companies -- one old, one new -- demonstrate a possible future for the aerospace parts industry. Yes, there will be robots.
- 'Harry Potter' audiobook narrator Stephen Fry said AI was used to steal his voice, and warned that convincing deepfake videos of celebrities will be nextwww.businessinsider.com 'Harry Potter' audiobook narrator Stephen Fry said AI was used to steal his voice, and warned that convincing deepfake videos of celebrities will be next
Stephen Fry said he was "shocked" when AI cloned his voice because it could "have me read anything from a call to storm Parliament to hard porn."
Stephen Fry said he was "shocked" when AI cloned his voice because it could "have me read anything from a call to storm Parliament to hard porn."
- Centaurs and Cyborgs on the Jagged Frontierwww.oneusefulthing.org Centaurs and Cyborgs on the Jagged Frontier
I think we have an answer on whether AIs will reshape work....
I think we have an answer on whether AIs will reshape work....
- I Asked AI to Generate a Superman Joker Fusion
YouTube Video
Click to view this content.
Subscribe for good luck - Most Popular videos - https://www.youtube.com/watch?v=72qmtr41iWU&list=UULP3MF3KCtKQkCudSvNpk3BOg#superman , #aiart , #joker \_Arti...
- Answering AI’s biggest questions requires an interdisciplinary approach | TechCrunchtechcrunch.com Answering AI’s biggest questions requires an interdisciplinary approach | TechCrunch
Ethical AI requires a deep understanding of what there is, what we want, what we think we know, and how intelligence unfolds.
Ethical AI requires a deep understanding of what there is, what we want, what we think we know, and how intelligence unfolds.
- Microsoft open sources EvoDiff, a novel protein-generating AI | TechCrunchtechcrunch.com Microsoft open sources EvoDiff, a novel protein-generating AI | TechCrunch
Microsoft has open sourced EvoDiff, an AI system and framework that can generate proteins without needing a protein sequence.
Microsoft has open sourced EvoDiff, an AI system and framework that can generate proteins without needing a protein sequence.
- Superorder raises $10M to help restaurants maintain their online presence | TechCrunchtechcrunch.com Superorder raises $10M to help restaurants maintain their online presence | TechCrunch
Superorder, a startup developing a platform to help restaurants maintain their online presence, has raised $10 million in a funding round.
Superorder, a startup developing a platform to help restaurants maintain their online presence, has raised $10 million in a funding round.
- Teachers Are Going All In on Generative AIwww.wired.com Teachers Are Going All In on Generative AI
Surveys suggest teachers use generative AI more than students, to create lesson plans or more interesting word problems. Educators say it can save valuable time but must be used carefully.
Surveys suggest teachers use generative AI more than students, to create lesson plans or more interesting word problems. Educators say it can save valuable time but must be used carefully.
- Google nears release of AI software Gemini, The Information reportswww.reuters.com Google nears release of AI software Gemini, The Information reports
Alphabet's Google has given a small group of companies access to an early version of Gemini, its conversational artificial intelligence software, The Information reported on Thursday, citing people familiar with the matter.
Alphabet's Google has given a small group of companies access to an early version of Gemini, its conversational artificial intelligence software, The Information reported on Thursday, citing people familiar with the matter.
- 'Dr. Google' meets its match in Dr. ChatGPT
In recent research AI has done a credible job at diagnosing health complaints. But should consumers trust unregulated bots with their health care? Doctors see trouble brewing.
- GitLab: Developers view AI as 'essential' despite concernswww.artificialintelligence-news.com GitLab: Developers view AI as 'essential' despite concerns
A survey by GitLab has shed light on the views of developers on the landscape of AI in software development.
A survey by GitLab has shed light on the views of developers on the landscape of AI in software development.
- UK's AI ecosystem to hit £2.4T by 2027, third in global racewww.artificialintelligence-news.com UK's AI ecosystem to hit £2.4T by 2027, third in global race
The UK's AI economy is soaring, valued at an impressive £1.36 trillion ($1.7 trillion) and showing no signs of slowing.
The UK's AI economy is soaring, valued at an impressive £1.36 trillion ($1.7 trillion) and showing no signs of slowing.
- UK government outlines AI Safety Summit planswww.artificialintelligence-news.com UK government outlines AI Safety Summit plans
The major event – set to be held at Bletchley Park, home of Alan Turing and other Allied codebreakers during the Second World War – aims to address the pressing challenges and opportunities presented by AI development on both national and international scales.
The major event – set to be held at Bletchley Park, home of Alan Turing and other Allied codebreakers during the Second World War – aims to address the pressing challenges and opportunities presented by AI development on both national and international scales.
- This driverless car company is using chatbots to make its vehicles smarterwww.technologyreview.com This driverless car company is using chatbots to make its vehicles smarter
Large language models are the next big thing for robotics, making cars and other robots quicker to train and easier to control (if you trust them).
Large language models are the next big thing for robotics, making cars and other robots quicker to train and easier to control (if you trust them).
- If you worry about humanity, you should be more scared of humans than of AIthebulletin.org If you worry about humanity, you should be more scared of humans than of AI
Advances in artificial intelligence have prompted extensive public concern about its capacity to contribute to the spread of misinformation, bias, and cybersecurity breaches—and its potential existential threat to humanity. But, if anything, AI can aid human beings in making decisions aimed at impro...
Advances in artificial intelligence have prompted extensive public concern about its capacity to contribute to the spread of misinformation, bias, and cybersecurity breaches—and its potential existential threat to humanity. But, if anything, AI can aid human beings in making decisions aimed at improving social equality, safety, productivity—and mitigate some existential threats.