[HN Gopher] Titans: Learning to Memorize at Test Time
       ___________________________________________________________________
        
       Titans: Learning to Memorize at Test Time
        
       Author : birriel
       Score  : 97 points
       Date   : 2025-01-13 20:11 UTC (4 days ago)
        
 (HTM) web link (arxiv.org)
 (TXT) w3m dump (arxiv.org)
        
       | cs702 wrote:
       | Interesting. I like the idea of a meta-mechanism that learns to
       | update an associative memory based on how surprising the data is.
       | The other stuff, reading memory via keys and values and
       | selectively erasing it with gating, look pretty conventional on a
       | first glance. Thank you for sharing this on HN. I've added it to
       | my reading list.
       | 
       | EDIT: I'm reminded of this other type of associative memory:
       | https://github.com/glassroom/heinsen_routing. The idea there is
       | to compute a mixture of memories that best predicts the given
       | input sequence. Quite frankly, I don't remember how the whole
       | thing works, but I do remember that it works. It's been a while
       | since I used it, so YMMV. In any case, it may be of interest to
       | you.
        
         | testfoo11111111 wrote:
         | there's nothing "pretty conventional" about a neural memory
         | mechanism that comes along with such solid evidence of
         | scalability and appealing performance characteristics.
         | 
         | If neural memory was conventional, GPT4o's memory wouldn't be
         | stored as plain text and prepended to prompts.
         | 
         | This paper reminds me of the Switch Transformer paper; e.g.
         | solidifying, expanding on, and proving out an area of research
         | that may well have a big impact on leading LLMs and the SOTA in
         | AI.
         | 
         | Agreed the concept of surprise is very cool.
        
           | pizza wrote:
           | There definitely is precedent - any parallelizably-decodable
           | CABAC-derived neural compression algorithm basically has a
           | flavor of this idea at its heart - intersperse statistical
           | state throughout your token stream so you can decouple
           | novelty in your state space on the fly.
           | 
           | Taken to its extreme where the 'memory' is descriptive enough
           | to deterministically control the decoding you get parallelism
           | over the sequence for free as a consequence of the
           | associativity.
           | 
           | Similar techniques are used in making video compression
           | algorithms robust enough for low latency reconnection in
           | online streaming in poor/changing network conditions, or
           | making it possible to decompress JPEGs at >1GBps in parallel
           | by exploiting the presence of 'RESET' tokens that indicate
           | independent/novel substreams.
           | 
           | That said, I do agree that this is definitely a great paper
           | and contribution to language models though!
        
           | cma wrote:
           | 1991
           | 
           | > Each RNN tries to solve the pretext task of predicting its
           | next input, sending only unexpected inputs to the next RNN
           | above. This greatly facilitates downstream supervised deep
           | learning such as sequence classification. By 1993, the
           | approach solved problems of depth 1000 (requiring 1000
           | subsequent computational stages/layers--the more such stages,
           | the deeper the learning). A variant collapses the hierarchy
           | into a single deep net. It uses a so-called conscious chunker
           | RNN which attends to unexpected events that surprise a lower-
           | level so-called subconscious automatiser RNN. The chunker
           | learns to understand the surprising events by predicting
           | them. The automatiser uses my neural knowledge distillation
           | procedure of 1991 [UN0-UN2] to compress and absorb the
           | formerly conscious insights and behaviours of the chunker,
           | thus making them subconscious. The systems of 1991 allowed
           | for much deeper learning than previous methods.
           | 
           | https://people.idsia.ch/~juergen/very-deep-
           | learning-1991.htm...
        
             | HarHarVeryFunny wrote:
             | It's unfortunate that Schmidhuber has both made many
             | seminal contributions to the field, but also engages in
             | "retroactive flag planting" whereby he claims credit for
             | any current successes that are remotely related to anything
             | he has worked on, even if only in terms of hand-wavy
             | problem approach rather than actually building upon his own
             | work.
             | 
             | It's obvious that things like memory, on various timescales
             | (incl. working), selective attention, surprise (i.e.
             | prediction failure) as a learning/memorization signal are
             | going to be part of any AGI solution, but the question is
             | how do you combine and realize these functionalities into
             | an actual cognitive architecture?
             | 
             | Schmidhuber (or in this case you, on his behalf!)
             | effectively saying "I worked on that problem, years ago" is
             | irrelevant. He also worked on LSTMs, which learned to
             | memorize and forget, and the reference section of the
             | "Titans" paper leads to many more recent attempts -
             | different proposed architectures - addressing the same
             | problems around (broadly speaking) learning how best to use
             | working memory. Lots of people suggesting alternatives, but
             | it would seem no compelling solution that has been
             | published.
             | 
             | If it's one of the commercial frontier model labs that does
             | discover the next piece of the architectural puzzle in
             | moving beyond transformers towards AGI, I very much doubt
             | they'll be in any hurry to publish it!
        
               | cma wrote:
               | "I like the idea of a meta-mechanism that learns to
               | update an associative memory based on how surprising the
               | data is."
               | 
               | Just pointing out that that idea was in some of
               | Schmidhuber's earlier work.
               | 
               | > Schmidhuber (or in this case you, on his behalf!)
               | effectively saying "I worked on that problem, years ago"
               | is irrelevant.
               | 
               | Ok. People do read his work and get ideas from it even if
               | this didn't necessarily. He had a lot of good stuff.
               | 
               | > but the question is how do you combine and realize
               | these functionalities into an actual cognitive
               | architecture?
               | 
               | I believe Schmidhuber gave one at the time?
        
               | sdenton4 wrote:
               | Does it work out-of-the-box today?
               | 
               | Execution is what matters. We can smoke a blunt and have
               | some nice sounding ideas, but building something that
               | works on data at scale is what actually counts.
        
               | cma wrote:
               | I think it's widely agreed a lot of useful stuff came out
               | of Schmidhubers lab. The example I gave was one of the
               | first things that scaled in lots of ways especially in
               | depth, and it shares some characteristics with this. I
               | doubt it outperforms this Titan architecture or is
               | equivalent. That's not the same as him just putting out
               | random ideas while high.
        
           | Xmd5a wrote:
           | >the concept of surprise is very cool
           | 
           | Then you may be interested in Simplicity Theory:
           | 
           | https://simplicitytheory.telecom-paris.fr/
           | Relevant situations are unexpected         Relevant features
           | generate compression         A situation or event is relevant
           | if it is unexpected.         This means that it is simpler to
           | describe than to generate.
           | 
           | In particular this recent paper:
           | 
           | >Unexpectedness and Bayes' Rule
           | 
           | >A great number of methods and of accounts of rationality
           | consider at their foundations some form of Bayesian
           | inference. Yet, Bayes' rule, because it relies upon
           | probability theory, requires specific axioms to hold (e.g. a
           | measurable space of events). This short document hypothesizes
           | that Bayes' rule can be seen as a specific instance of a more
           | general inferential template, that can be expressed also in
           | terms of algorithmic complexities, namely through the measure
           | of unexpectedness proposed by Simplicity Theory.
           | 
           | Source:
           | https://cifma.github.io/Papers-2021/CIFMA_2021_paper_13.pdf
        
             | Vampiero wrote:
             | It's hard to take it seriously when every single paper on
             | the subject is from one guy
        
       | gwern wrote:
       | Duplicate: https://news.ycombinator.com/item?id=42718166
        
         | cs702 wrote:
         | That page has more substantive comments. Your comment should be
         | at the top here.
        
         | birriel wrote:
         | OP here. I made this submission 3 days ago. The thread you're
         | referencing was posted yesterday, using the same exact link.
         | 
         | HN needs to do better.
        
       ___________________________________________________________________
       (page generated 2025-01-17 23:01 UTC)