Pat Inc的封面图片
Pat Inc

Pat Inc

IT 服务与咨询

Palo Alto,California 542 位关注者

A scientific breakthrough in NLU for conversational AI. Sign up for early access on our website.

关于我们

Pat is a Conversational AI platform that provides meaning - the missing piece - to create human-like interactions with machines. How? By mapping the language to the meaning behind words such as context, synonyms, and previous conversations where reasoning and 'common sense' are by-products. A scientific breakthrough in Natural Language Understanding (NLU), Pat enables human-like accuracy for conversation, is language independent and domain agnostic, and has easier integration and ongoing management. Sign up for early access: https://pat.ai/

网站
https://www.pat.ai
所属行业
IT 服务与咨询
规模
2-10 人
总部
Palo Alto,California
类型
私人持股
创立
2015
领域
natural language understanding、machine intelligence、artificial intelligence、voice assistants、chatbots、intelligent assistants、NLU、natural language和conversational AI

地点

Pat Inc员工

动态

  • Pat Inc转发了

    An excerpt from today's substack discussion about the next generation of AI. Will AGI come from text generators or will it follow the plan to emulate human beings? In the video clip, a human would react in about a quarter of a second in either scenario. How would a text-based system do this without incorporating robotics and its associated sensory apparatus. When you look at well funded LLM companies, they seem to be deviating from the meaning of AGI to avoid the difficult parts of AI. For this reason I argue that today's LLMs are not going to be on the path to AGI because they are not focused on the building blocks of AI - animal level brain capabilities that support real 'strong AI'.

  • Pat Inc转发了

    查看Beth Carey的档案

    Founder and Director @ Pat Inc | The new language interface

    The ultimate sustainable solution has the efficiency and accuracy of the human brain. When 'hallucinations' are a bug not a feature as for enterprise software, there is a better way forward. But how do we emulate the brain to achieve language and vision comprehension with those benefits? That better way forward is brain based natural language understanding from the cognitive sciences. John Ball writes about that gold standard language model required for the trustworthy language interface in 'How to solve AI with Our Brain' (link below in comments). #PatomTheory #RRGLinguistics

    查看Boris Gamazaychikov的档案

    Head of AI Sustainability @ Salesforce

    Lack of transparency is a fundamental challenge to AI sustainability. That's why we're introducing ???? ???????????? ?????????? – a standardized way to measure and compare the energy efficiency of AI models. What’s included? ?? Interactive Leaderboard – data from hundreds of models already scored https://lnkd.in/gjvu5BEc ?? Submission Portal – score both open and closed-source models https://lnkd.in/gb3z4Hbv ??? Label Generator – share results far and wide https://lnkd.in/gJaSHfbJ ?? Documentation + FAQ – learn more about the approach https://lnkd.in/g3uEk_yA ?? Community Hub – to discuss and connect https://lnkd.in/g6-ceFiG This is just the beginning! We need your help to make this an industry standard and get more AI models disclosing their energy efficiency. Reach out and check out the FAQ to learn how! Huge congrats to everyone who made this happen! Dr. Sasha Luccioni Clem Delangue ?? Yacine Jernite Régis Pierrard Sara Hooker Silvio Savarese Itai Asseo Denise Pérez Michael Weimann Hannah Downey Julie Ravillon Laurent Monjole Serena Ingre

  • Pat Inc转发了

    查看Beth Carey的档案

    Founder and Director @ Pat Inc | The new language interface

    For those interested in following the path John Ball wrote about last week to bring his technology to life, I will be posting about it too. You might be wondering how enabling language comprehension for machines helps language learners. Well at the core of his system, is a linguistic framework. Based on 'meaning' and extensive research of the world's languages, he used it to teach machines language in a machine readable format. When language educator, author and performance coach Chris Lonsdale saw it, he realized it would have major implications for second language learning because it was, well, brain-like. So more will be revealed about this intersection of John's cognitive science breakthroughs and Chris' language learning pedagogy as we build the first application combining the two. For now, here is a summary of Chris' original TEDx talk with over 35 million views on YouTube alone!

  • Pat Inc转发了

    This is my unedited reaction video interacting with LLMs - Leo from Brave and CoPilot from Microsoft. You can see how human context is ignored and, instead, non-human associations are assumed. It is ***funny***!! I tried to stay serious. You can see how assigning a name skews results based on the statistical uses of that name, even though a name is arbitrary in a human language until resolved to a specific case. Early on, the word 'promise' is misrepresented as 'convince,' that even fooled me for a while until I spotted that subtle but catastrophic error. I will show you the correct meaning from my language tool in a future video. There is very strong linguistic science behind how context works. ICG, GCG and CoU describe different types of context in human conversation, for example. I'll do a more detailed analysis in my substack for those who want to understand why the failures took place in the video. Summary: LLMs don't emulate human context which makes them very unlike humans with language.

  • Pat Inc转发了

    LLMs make stuff up. How do we fix that? The attached demonstration shows the magnitude of problem that comes from a statistical model of a knowledge base - best fixed by changing to a meaning-based and accurate model. Patching the model is the wrong approach since the training today is too expensive and a real solution with normal programming is demonstrably cheaper and more accurate.

  • Pat Inc转发了

    Why are some predictions accurate and others fail by generations. Most AI predictions have been inaccurate because they need scientific breakthroughs. Engineering breakthroughs as needed in building rockets and computers are easier because the science is already solved. Today I look at those questions as well as explaining the difference between science and engineering.

  • 查看Pat Inc的组织主页

    542 位关注者

    查看Beth Carey的档案

    Founder and Director @ Pat Inc | The new language interface

    Gary Marcus has long stood up for the 'science' because for cognitive scientists, they know the fallacy of scaling computation to emulate brain functions like language and vision. "The thing is,?in the long term, science isn’t majority rule. In the end, the truth generally outs. Alchemy had a good run, but it got replaced by chemistry. The truth is that scaling is running out, and that truth is, at last coming out" It's been an unpopular message. Gary describes the vilification by proponents of the current scaling paradigm in his Newsletter. Where as generative AI is considered an off ramp by some to AI, there is an alternative where machines *can* be controlled by the trustworthy multi-model interface with language and vision for society's benefit. It just won't arrive without brain emulation enabled by cognitive science first. #beyondLLMs #deeplearning #linguistics #patomtheory

  • 查看Pat Inc的组织主页

    542 位关注者

    查看Beth Carey的档案

    Founder and Director @ Pat Inc | The new language interface

    Ok - I have an update. Thank you so much for your support globally for the imminent release of John Ball 's book "How to Solve AI with our Brain" ?? After some sleepless nights, our book is in its final stages of Amazon review ?? YAY and a new link has been issued. For those of us who have already pre-ordered, (yes, me included ?? ) you were not charged but will need to resubmit your pre-order on the new link below to secure an order. I've just done it, so if you have any questions, please let me know. I'm sorry for the double handling but we had a delay on our end which led to a cancellation of the original link. Since we're only hours away from the big day, if you haven't already, be among the first to dive into 'How to solve AI with our Brain'. Not only is it timely, but I am looking forward to the clarity it will bring at a time when there is so much misinformation, especially in 'AI'. Aimed at the general audience, from high school to any age group who just wants to learn, I know you'll find fascinating titbits. I'm looking as forward to the conversation as I am to the controversy as the book will inform, explain and illuminate backed up with new cognitive science, fun narrative, diagrams and cute line art. While many in the industry are talking about 'beyond LLMs' this is it! Dubbed 'the dawn of Meaning AI' John Ball takes the reader on a journey from the evolution of the animal and human brain, his fascination since a child with human language and vision for computers, to coming up with a brain model which can be emulated on a machine. A virtuous circle. ? The book is a chance to open up the narrative beyond the one saturating the press by tech giants. Yes, there *is* a better option, you just may not have heard about it yet ?? . Spoiler Alert: the brain doesn't process like a computer! Reach out to me if you have any questions, or comments when you get a copy! #PatomTheory #RRGLinguistics #NoMoreTrainingData

  • Pat Inc转发了

    查看Pat Inc的组织主页

    542 位关注者

    查看NewsRamp?的组织主页

    1,448 位关注者

    John Ball, a renowned AI luminary, is set to release his groundbreaking book, 'How to Solve AI with Our Brain.' Published by Pat Labs Pty Ltd, this insightful work exposes the limitations of computational AI and uncovers cognitive science breakthroughs that could meaningfully advance the field. 'While many consumers encounter AI primarily through computational systems, this book also addresses the higher standards for certainty and reliability that enterprise software users and producers demand,' says Dr. Neal Sample, Chief Information Officer at Walgreens Boots Alliance. Ball empowers readers to understand the scientific foundations that will shape the future of AI, moving beyond the flawed promises of current computational approaches. The book is a must-read for anyone interested in the intersection of cognitive science and artificial intelligence. Copies of 'How to Solve AI with Our Brain' can be ordered from November 1st, 2024, as an e-book or print edition. #ArtificialIntelligence #CognitiveScience #ReliableAI

相似主页

查看职位

融资

Pat Inc 共 1 轮

上一轮

种子轮

US$2,500,000.00

Crunchbase 上查看更多信息