Meta tags:

Headings (most frequently used words):

daniel, filan, about, me, papers, bit, out, of, date, sorry,

Text of the page (most frequently used words):
and (19), the (19), that (10), for (9), with (8), where (6), can (6), agents (5), networks (5), you (5), priors (4), them (4), planning (4), reinforcement (4), learning (4), have (4), also (4), research (4), podcast (4), about (4), marcus (3), hutter (3), presented (3), speed (3), complexity (3), programs (3), time (3), lead (3), author (3), how (3), arxiv (3), write (3), models (3), inverse (3), use (3), neurons (3), into (3), clusterability (3), are (3), read (3), here (3), was (3), axrp (3), filan (3), 2016 (2), complex (2), paper (2), their (2), modifications (2), they (2), from (2), web (2), book (2), probabilistic (2), topics (2), such (2), 2018 (2), more (2), one (2), task (2), stuart (2), russell (2), introduces (2), groups (2), edges (2), between (2), than (2), this (2), neural (2), did (2), year (2), thesis (2), 2024 (2), called (2), episodes (2), listen (2), youtube (2), risk (2), blog (2), metr (2), daniel (2), jan, leike, aistats, discussion, say, over, infinite, sequences, bits, penalise, strings, measured, length, produce, string, those, take, run, builds, off, jürgen, schmidhuber, defining, his, prior, original, jmlr, loss, bounds, tom, everitt, mayank, daswani, agi, winner, kurzweil, prize, best, discusses, modify, source, code, predict, result, these, define, don, make, stop, optimising, what, originally, told, optimise, self, modification, policy, utility, function, rational, owain, evans, andreas, stuhlmüller, john, salvatier, explaining, programming, language, covers, inference, mdps, hyperbolic, discounting, myopic, multi, agent, webppl, website, modeling, chris, cundy, goalsrl, held, jointly, icml, ijcai, aamas, advocates, hierarchical, humans, realistic, tasks, showing, perform, comparably, state, art, exploring, hierarchy, aware, stephen, casper, shlomi, hod, cody, wild, andrew, critch, dividing, network, same, group, higher, weight, different, implements, using, graph, clustering, refers, divisibility, shows, many, conditions, trained, pruning, dropout, clusterable, weights, were, randomly, permuted, method, regularizing, bibtex, papers, bit, out, date, sorry, undergrad, australian, national, university, studying, theory, mathematics, theoretical, physics, honours, similar, master, degree, lasting, under, resource, bounded, based, completed, phd, berkeley, supervised, structure, representation, mid, late, 2025, senior, manager, chatted, scholars, hopefully, turned, cool, resarchers, specifically, working, alignment, interpretability, governance, managed, other, managers, mats, interested, our, limited, resources, most, good, world, sometimes, reasons, described, immanuel, kant, bryan, caplan, bet, things, effective, altruism, addition, another, talk, people, whatever, want, available, wherever, podcasts, cabinet, which, short, searching, your, favourite, app, alteratively, transcripts, interest, there, posts, puzzles, math, forecasting, until, recently, member, technical, staff, helped, manage, evaluations, ran, fed, frontier, report, github, pgp, public, key, email,

Text of the page (random words):
daniel filan daniel filan blog email pgp public key cv github about me until recently i was a member of technical staff at metr where i helped manage evaluations metr ran that fed into the frontier risk report i have a blog where i write about topics of interest to me as of the time i write this there are posts about forecasting math and puzzles i also have a podcast it s called axrp which is short for the ai x risk research podcast you can listen to episodes on youtube or by searching axrp in your favourite podcast app alteratively you can read transcripts here in addition to axrp i have another podcast called the filan cabinet where i talk to people about whatever i want episodes are available on youtube or wherever you listen to podcasts i m interested in effective altruism how we can use our limited resources to do the most good in the world i also sometimes bet on things for reasons described by bryan caplan and immanuel kant from mid 2024 to late 2025 i was a senior research manager at mats where i chatted with scholars and hopefully turned them into cool resarchers specifically working on ai alignment interpretability and or governance and also managed other research managers i completed my phd in ai at uc berkeley in 2024 where i was supervised by stuart russell you can read my thesis structure and representation in neural networks here i did my undergrad at the australian national university studying the theory of reinforcement learning mathematics and theoretical physics i did my honours year similar to a research master s degree lasting one year under marcus hutter you can read my thesis resource bounded complexity based priors for agents here papers a bit out of date sorry bibtex clusterability in neural networks arxiv with stephen casper shlomi hod cody wild andrew critch and stuart russell introduces the task of dividing the neurons of a network into groups such that edges between neurons in the same group have higher weight than edges between neurons in different groups implements this using graph clustering so clusterability refers to the divisibility of networks shows that in many conditions networks trained with pruning and or dropout are more clusterable than if their weights were randomly permuted also introduces a method of regularizing networks for clusterability exploring hierarchy aware inverse reinforcement learning arxiv with chris cundy lead author presented at goalsrl 2018 held jointly at icml ijcai and aamas 2018 advocates for the use of hierarchical planning models of humans for use in inverse reinforcement learning as more realistic for complex tasks showing that in one task they perform comparably to state of the art models modeling agents with probabilistic programs website with owain evans lead author andreas stuhlmüller and john salvatier web book a web book explaining how to write models of agents in the webppl probabilistic programming language covers topics such as planning as inference po mdps inverse reinforcement learning hyperbolic discounting myopic planning and multi agent planning self modification of policy and utility function in rational agents arxiv with tom everitt lead author mayank daswani and marcus hutter presented at agi 2016 winner of the kurzweil prize for best paper discusses agents that can modify their source code and predict the result of these modifications and how to define them so that they don t make modifications that stop them from optimising what we originally told them to optimise loss bounds and time complexity for speed priors jmlr with jan leike and marcus hutter presented at aistats 2016 a discussion of speed priors that is to say priors over infinite sequences of bits that penalise complex strings where complexity is measured by the length of programs that produce a string and the time those programs take to run builds off jürgen schmidhuber s original paper defining his speed prior