Lifeboat Foundation: Safeguarding Humanity
Toggle light / dark theme

Blog – Page 759

As Shumer told VentureBeat over DM: “I’ve been thinking about this idea for months now. LLMs hallucinate, but they can’t course-correct. What would happen if you taught an LLM how to recognize and fix its own mistakes?”

Hence the name, “Reflection” — a model that can reflect on its generated text and assess its accuracy before delivering it as outputs to the user.

The model’s advantage lies in a technique called reflection tuning, which allows it to detect errors in its own reasoning and correct them before finalizing a response.

Read more | >

Why a blog post and not a proper response in a philosophy journal? My very first journal submission is still in the review process, close to two months later, for one. Secondly, blogging allows me to be pedantic, to be human, that is, to express frustration, to show anger, to be candid; in other words, blogging allows me to be myself. Probably of highest priority is the fact that I do not want my first publication in the philosophy of mind to be a response. I want to eventually outline my own theory of consciousness, which is strongly hinted at here, and I prefer for that to be my first contribution to the philosophy of mind. I do not find panpsychism convincing and I think there is another theory of consciousness, similar to panpsychism in ways, that is much more cogent. I have outlined some qualms I have with panpsychism before; to people new to the blog, you can read here.

Read more | >

Researchers from the University of Pisa developed a quantum subroutine to improve matrix multiplication for AI and machine learning applications.

When you multiply two large matrices—this is a common task in fields like machine learning, but it can be time-consuming, even for powerful computers…

In a recent study published in IEEE Access, a team of researchers from the University of Pisa introduced a quantum subroutine designed to streamline matrix multiplication. This subroutine is a new feature in the toolbox of matrix multiplication that could improve computational efficiency, particularly in applications like machine learning and data processing.

It’s A Matrix World And We’re Just Living In It

Continue reading “Matrix Re-Reloaded: Quantum Subroutine Improves Efficiency of Matrix Multiplication for AI and Machine Learning Applications” | >

Imagine you’re sitting across from a friend, having a conversation.

I’m a die-hard Beach Boys fan. In one of their most famous songs, they sing about “pickin’ up good vibrations” from a girl. We’ve all felt those “good vibes” when we’re connecting with someone new. I used to think that feeling was a mysterious, mystical experience — something I couldn’t fully explain that bonded me with some friends and strangers more easily than others.

It turns out that “good vibes” aren’t as mysterious as I thought.

Pioneering neuroscientists have begun investigating how the brain works when we are interact ing with others — a technique they call hyperscanning. Neuroscientists have been using existing scanning methods, like MRI and EEG, to monitor the brain activity of two or more people as they do something together: for example, performing music, learning a poem, or having a conversation.

Continue reading “Picking up good vibes from a stranger? That might be your neurons aligning” | >

Learning and a spectrum of other behavioral competencies allow organisms to rapidly adapt to dynamically changing environmental variations. The emerging field of diverse intelligence seeks to understand what systems, besides ones with complex brains, exhibit these capacities. Here, we tested predictions of a general computational framework based on the free energy principle in neuroscience but applied to aneural biological process as established previously, by demonstrating and manipulating pattern recognition in a simple aneural organism, the green algae Volvox. Our studies of the adaptive photoresponse in Volvox reveal that aneural organisms can distinguish between patterned and randomized inputs and indicate how this is achieved mechanistically.

Read more | >