Truth Tide TV UNSEALED Epstein Case Files
menu
home Home analytics Reports article Articles auto_stories Narratives mail Email description Documents videocam Videos search Search
policy Investigate expand_more
inbox Inbox 74547 send Sent 28705 label All Mail 74547 attach_file Attachments 1907 topic Topics
People
Jeffrey Epstein person
Ghislaine Maxwell person
Bill Clinton person
Alan Dershowitz person
Elon Musk person
Bill Gates person
Ehud Barak person
Reid Hoffman person
Peter Thiel person
Larry Summers person
Prince Andrew person
Steve Bannon person
Masha Bucher person
Jason Calcanis
Michael Wolff person
Noam Chomsky person
Tom Pritzker person
Al Seckel person
Kimbal Musk person
Karyna Shuliak person
Deepak Chopra person
Ken Starr person
Peter Attia person
Jeremy Rubin person
Neri Oxman person
Marvin Minsky person
Lawrence Krauss person
Seth Lloyd person
Boris Nikolic person
Jean Luc Brunel person
Lesley Groff person
Sarah Kellen person
Nadia Marcinkova person
Darren Indyke person
Mark Epstein person
Emad Hanna person
Joscha Bach person
Rich Kahn person
Cecelia Steen
John Amerling person
Sultan Bin Sulayem person
Matthew Hitzik
Peter Mandelson person
groups People directory
74547 threads 209740 messages
Homechevron_right Emailchevron_right Mobjects and Replisomes
arrow_back

Mobjects and Replisomes

1 message picture_as_pdf Source PDF
J
Joscha Bach Jun 5, 2015 3:36 AM
To
Jeffrey EpsteinMartin Nowak

Dear Jeffrey,

thank you for an inspiring afternoon! I believe that the 'mobjects' that you have been talking about are going to become a very hot topic. They must be above the level of neurons, and below, or orthogonal to most mechanisms of language. Mathematically, we can probably treat them as directed hypergraphs with typed links, but we rarely go to that level of description. I guess that Carnap and Tarski delivered the first incarnations of the modern theories on them, and every comprehensive theory of thinking, planning and language has to account for them, and yet our understanding is preliminary. In AI, we got Schank's schemas, scripts and MOPs, which resulted in Minsky's descriptions of frames, which in turn probably informed Danny Hillis' work on concept net (which is now the foundation of Google's knowledge graph). Jeff Hinton is currently talking about "capsules" that are artificial neural implementations of mobjects, and Markram has been trying to go after them by modeling neural columns. In MicroPsi, I call them concept nodes, in OpenCog, they are the elements of the AtomTable, and John Anderson talks about chunks. I would be very interested to learn about Noam's deal; I sense that he is mostly familiar with approaches like Winston's, who uses symbolic approximations?

Hope that it continues soon!

Joscha

Dear Martin,

perhaps you will like this lesser-known Bjark video (I discovered it Sunday at the MoMa). It is part of her "biophilia" album, and travels in a stepwise order-of-magnitude enlargement into the molecular mechanisms of cellular replication; the song is inspired by the thought of individuals representing parts of a lineage, both on the cellular and the organismic level. To me, the animation is one of the most beautiful and enlightening ones that I have seen on the topic, especially considering its short duration.

https://www.youtube.com/watch?v=Wa1A0pPc-ik

Nucleosomes start at about 1:40, halfway through the video, a groove protein travels along a strand of DNA, and beginning from 3:05, there is a phantastically animated replisome. I wished I could get the animator (Drew Berry) to do a similar thing that elucidates neural networks... Cheers,

Joscha

1419 files from the DOJ Epstein case media release. All files are public records from justice.gov.

Built by Truth Tide TV