Fun

News Feed - 2023-10-26 02:10:46

Tristan Greene6 hours agoResearchers in China developed a hallucination correction engine for AI modelsThe “Woodpecker” hallucination correction system can, ostensibly, be applied to any multimodal large language model, according to the research.980 Total views36 Total sharesListen to article 0:00NewsJoin us on social networksA team of scientists from the University of Science and Technology of China and Tencent’s YouTu Lab have developed a tool to combat “hallucination” by artificial intelligence (AI) models. 


Hallucination is the tendency for an AI model to generate outputs with a high level of confidence that don’t appear based on information present in its training data. This problem permeates large language model (LLM) research, and its effects can be seen in models such as OpenAI’s ChatGPT and Anthropic’s Claude.


The USTC/Tencent team developed a tool called “Woodpecker” that they claim is capable of correcting hallucinations in multimodal large language models (MLLMs). 


This subset of AI involves models such as GPT-4 (especially its visual variant, GPT-4V) and other systems that roll vision and/or other processing into the generative AI modality alongside text-based language modeling. 


According to the team’s preprint research paper, Woodpecker uses three separate AI models, apart from the MLLM being corrected for hallucinations, to perform hallucination correction. 


These include GPT-3.5 turbo, Grounding DINO and BLIP-2-FlanT5. Together, these models work as evaluators to identify hallucinations and instruct the model being corrected to regenerate its output in accordance with its data. In each of the above examples, an LLM hallucinates an incorrect answer (green background) to prompting (blue background). The corrected Woodpecker responses are shown with a red background. Source: Yin, et. al., 2023


To correct hallucinations, the AI models powering Woodpecker use a five-stage process that involves “key concept extraction, question formulation, visual knowledge validation, visual claim generation, and hallucination correction.”


Related:Humans and AI often prefer sycophantic chatbot answers to the truth — Study


The researchers claim these techniques provide additional transparency and “a 30.66%/24.33% improvement in accuracy over the baseline MiniGPT-4/mPLUG-Owl.” They evaluated numerous “off the shelf” MLLMs using their method and concluded that Woodpecker could be “easily integrated into other MLLMs.”# Science# AI# ChatGPTAdd reactionAdd reactionRead moreAI a powerful tool for devs to change gaming, says former Google gaming headBlockchain companies are creating AI chatbots to help developersOpenAI partners with G42 in Dubai, eyeing Middle East expansion

News Feed

VC Roundup: AI-blockchain solutions winning investors’ hearts
Ana Paula Pereira5 hours agoVC Roundup: AI-blockchain solutions winning investors’ heartsUtila, Synnax, Sahara, TON Foundation, UXLINK, and TEN are among the startups securing funding in the first half of March.4230 To
SingularityNET and Filecoin partner for AI and DePIN
Josh O"Sullivan11 hours agoSingularityNET and Filecoin partner for AI and DePINSingularityNET and Filecoin collaborate to integrate decentralized AI and DePIN, emphasizing ethical AI development and decentralized infrast
Family of Popular NFT Projects Launch Collaborative Initiative and Genesis Collection
Family of Popular NFT Projects Launch Collaborative Initiative and Genesis Collection On Tuesday, a new Draper Goren Holm incubated company called NFT.FAMILY has announced the laun
Anti-crypto SEC commissioner’s term is up in 41 days — Will she be replaced?
Turner Wright1 hour agoAnti-crypto SEC commissioner’s term is up in 41 days — Will she be replaced?Caroline Crenshaw may continue to serve at the SEC for an additional 18 months after her term ends in June if the Whi
Solana Picture Bigger Than $420: Analyst Predicts 140% Surge To New ATHs
Reason to trust Strict editorial policy that focuses on accuracy, relevance, and impartiality Created by industry experts and meticulously reviewed The highest standards in reporting and pu
EasyA announces 1 million devs in ecosystem
Vince Quill1 hour agoEasyA announces 1 million devs in ecosystemThe Web3 educational platform also announced a Polkadot education initiative to help onboard new developers to the blockchain network.153 Total viewsListen
Polygon Reveals Supernets Project to Bolster Web3 Adoption, Pledges $100 Million to Early Users
Polygon Reveals Supernets Project to Bolster Web3 Adoption, Pledges $100 Million to Early Users On April 22, the layer two (L2) scaling network Polygon introduced the new Supernets
Rakesh Upadhyay5 hours agoPrice analysis 9/6: BTC, ETH, BNB, XRP, ADA, DOGE, SOL, TON, DOT, MATICBitcoin and several major altcoins continue to show weakness, but a few data points do favor bulls in the short term.973 To
Seamless Protocol introduces Integrated Liquidity Market on Lido
Ana Paula Pereira6 hours agoSeamless Protocol introduces Integrated Liquidity Market on LidoSeamless’ first Integrated Liquidity Market will enable stakers on Lido to automatically use a borrowing strategy and compound
Base overtakes Arbitrum for most active addresses so far this month
Brayden Lindrea4 hours agoBase overtakes Arbitrum for most active addresses so far this monthCoinbase layer-2 network Base has seen a strong start to April with active addresses outpacing Arbitrum.903 Total views31 Total
Hong Kong Company Onedegree to Offer Insurance for Cryptocurrencies
Hong Kong Company Onedegree to Offer Insurance for Cryptocurrencies Onedegree, a virtual insurer based in Hong Kong, will provide insurance services for digital assets as part of a
Latam Common Currency to Focus on Settlements Among Mercosur and BRICS Countries
Latam Common Currency to Focus on Settlements Among Mercosur and BRICS Countries President of Brazil Luis Inacio Lula Da Silva clarified the scope and reach of the common currency