Lifeboat Foundation

One key aspect of intelligence is the ability to quickly learn how to perform a new task when given a brief instruction. For instance, a child may recognise real animals at the zoo after seeing a few pictures of the animals in a book, despite any differences between the two. But for a typical visual model to learn a new task, it must be trained on tens of thousands of examples specifically labelled for that task. If the goal is to count and identify animals in an image, as in “three zebras”, one would have to collect thousands of images and annotate each image with their quantity and species. This process is inefficient, expensive, and resource-intensive, requiring large amounts of annotated data and the need to train a new model each time it’s confronted with a new task. As part of DeepMind’s mission to solve intelligence, we’ve explored whether an alternative model could make this process easier and more efficient, given only limited task-specific information.

Today, in the preprint of our paper, we introduce Flamingo, a single visual language model (VLM) that sets a new state of the art in few-shot learning on a wide range of open-ended multimodal tasks. This means Flamingo can tackle a number of difficult problems with just a handful of task-specific examples (in a “few shots”), without any additional training required. Flamingo’s simple interface makes this possible, taking as input a prompt consisting of interleaved images, videos, and text and then output associated language.

Similar to the behaviour of large language models (LLMs), which can address a language task by processing examples of the task in their text prompt, Flamingo’s visual and text interface can steer the model towards solving a multimodal task. Given a few example pairs of visual inputs and expected text responses composed in Flamingo’s prompt, the model can be asked a question with a new image or video, and then generate an answer.

Earlier this month, we brought you the news that epigenetic reprogramming startup YouthBio Therapeutics had emerged from stealth. The company shed some light on its plans to develop epigenetic reprogramming therapies for age-related diseases by rejuvenating certain cells in our bodies. YouthBio aims to achieve this rejuvenation by developing gene therapies that enable partial cellular reprogramming – an area of longevity science that is now attracting significant commercial interest.

Longevity. Technology: Cellular reprogramming refers to the process of returning adult cells to a “pluripotent” state: blank, embryonic-like cells that can become any cell in the body. This reprogramming can be achieved using techniques based on the discovery of Yamanaka factors.

0:00 PeopleLens AI Helps The Blind.
1:40 Brain Fingerprints Detect Autism.
4:52 AI Predicts Cancer Tumor Regrowth.

Learn more about the future of decentralized AI here:
SingularityNET AGIX Website — https://singularitynet.io
Developer Documentation — https://dev.singularitynet.io/
Publish AI Services — https://publisher.singularitynet.io/
AGIX Community Telegram — https://t.me/singularitynet
AGIX Price Chat Telegram — https://t.me/AGIPriceTalk

Continue reading “AI News Timestamps” »

Planetary Protection — how strict should the guidelines be to protect our home planet, to protect other worlds?

For more info see:

Continue reading “Three Myths of Planetary Protection” »

Not even death itself will prevent this tech-savvy grandfather from meeting his future great-grandchildren.

Not even death will stop this tech-savvy grandfather from meeting his great-grandchildren.

Jerry Terrance, an 85-year-old grandfather from Los Angeles, California, has turned himself into a 3D hologram that will serve as a humanoid time capsule for future generations. According to the Daily Mail, Jerry’s ‘hologram twin’ will guide his two children and four grandchildren, as well as future great-grandchildren, through his family’s history, even after his death.

Continue reading “Grandpa Creates Hologram Twin For Future Grandkids Using VR” »

In many fields of technology, smaller is better, and machinery is now getting so tiny it’s measured in mere atoms. Researchers at the University of Erlangen–Nuremberg (FAU) in Germany have now developed what they claim are the world’s smallest working gear wheels.

Molecular machines and nanorobots could be extremely useful in the coming decades, helping to construct electronic components, transport drugs through the body, or manipulate individual cells or molecules.

To that end, scientists have developed nanoscale versions of many machine parts, such as motors, pistons, pumps, wrenches and propellers.

Use code ISAACARTHUR16 for up to 16 FREE MEALS + 3 Surprise Gifts across 6 HelloFresh boxes plus free shipping at https://bit.ly/3uRoKz7
As computers have improved at an accelerating rate for generations now, fears of some emergent super intelligent computer mind have grown, but is such a Technological Singularity Inevitable, and can we survive it?

Visit our Website: http://www.isaacarthur.net.
Support us on Patreon: https://www.patreon.com/IsaacArthur.
Support us on Subscribestar: https://www.subscribestar.com/isaac-arthur.
Facebook Group: https://www.facebook.com/groups/1583992725237264/
Reddit: https://www.reddit.com/r/IsaacArthur/
Twitter: https://twitter.com/Isaac_A_Arthur on Twitter and RT our future content.
SFIA Discord Server: https://discord.gg/53GAShE

Continue reading “Is a Technological Singularity Inevitable?” »

Scientists led by Dr. Manuel Serrano have observed symptoms of rejuvenation in the pancreas, liver, spleen, and blood of mice after applying one cycle of cell reprogramming.

To achieve this, the researchers have characterized rejuvenation by studying molecular marks in the DNA, gene expression, and cell metabolism. The study has been published in the journal Aging Cell.

Continue reading “Researchers Rejuvenate Mouse Organs Through Cellular Reprogramming” »

HPE has launched its Swarm Learning and Machine Learning Development System whilst Salesforce announced further integration with Slack, Tableau, and Mulesoft.