Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

TimesFM: Time Series Foundation Model for time-series forecasting

github.com

Post reply on HN
Print view View original

111–120 of 123 posts

Re: TimesFM: Time Series Foundation Model for time-series forecasting

#111

Post by nextaccountic »

itronitron wrote: post #68
whimsicalism wrote: post #17

I'm curious why we seem convinced that this is a task that is possible or something worthy of investigation. I've worked on language models since 2018, even then it was obvious why language was a useful and transferable task. I do not at all feel the same way about general univariate time series that could have any underlying process.

There was a paper written a while back that proved mathematically how you can correlate any time series with any other time series, thus vaporizing any perception of value gained by correlating time series (at least for those people that read the paper.) just wanted to share

Do you have a link to the paper?

Re: TimesFM: Time Series Foundation Model for time-series forecasting

#112

Post by pvorb »

amelius wrote: post #107
Xcelerate wrote: post #98

Earlier quoted context omitted.

Is Rentech the only group that genuinely manages to predict stock price? Seems like the very observation that it’s still possible would be enough motivation for other groups to catch up over such a long period. Also, the first realistic approximation of Solomonoff induction we achieve is going to be interesting because it will destroy the stock market.

Maybe that would be a good thing. I wouldn't mourn the destruction of the stock market as it's just a giant wealth-gap increasing casino. Trading has nothing to do with underlying value.

I fully agree. The stock market is just a giant machine that pulls money out of systems.

Re: TimesFM: Time Series Foundation Model for time-series forecasting

#113

Post by amelius »

icapybara wrote: post #101
Xcelerate wrote: post #98

Earlier quoted context omitted.

Is Rentech the only group that genuinely manages to predict stock price? Seems like the very observation that it’s still possible would be enough motivation for other groups to catch up over such a long period. Also, the first realistic approximation of Solomonoff induction we achieve is going to be interesting because it will destroy the stock market.

Agreed, if stock prices were predictable by some technical means, they would be quickly driven to unpredictability by people trading on those technical indicators.

And that would make the stock markets accessible to fewer people, further widening the wealth gap.

Re: TimesFM: Time Series Foundation Model for time-series forecasting

#114

Post by viraptor »

morkalork wrote: post #100
viraptor wrote: post #77

Earlier quoted context omitted.

The more complex the data is, the more you need. If your values are always 5, then you need only one data point.

If your values were always 5,you wouldn't use an LSTM to model it either. So presumably there's a threshold for when LSTM becomes practical and useful, no?

Sure, that was an extreme example. The point was that the minimum of data is 1 point, maximum is "all of it". It entirely depends on your use case.

Re: TimesFM: Time Series Foundation Model for time-series forecasting

#115

Post by Xcelerate »

belter wrote: post #106
Xcelerate wrote: post #98

Earlier quoted context omitted.

Is Rentech the only group that genuinely manages to predict stock price? Seems like the very observation that it’s still possible would be enough motivation for other groups to catch up over such a long period. Also, the first realistic approximation of Solomonoff induction we achieve is going to be interesting because it will destroy the stock market.

Rentech does not seem to be able to predict the stock market for their customers... "Jim Simons' Renaissance Technologies suffers $11 billion of client withdrawals in 7 months" - https://markets.businessinsider.com/news/stocks/jim-simons-r...

I’m referring to RenTech’s well known Medallion fund, which I believe is now only available internally to longtime employees. Even in the article you linked, it says this fund has “continued to shine”.

Re: TimesFM: Time Series Foundation Model for time-series forecasting

#116

Post by belter »

Xcelerate wrote: post #115
belter wrote: post #106

Earlier quoted context omitted.

Rentech does not seem to be able to predict the stock market for their customers... "Jim Simons' Renaissance Technologies suffers $11 billion of client withdrawals in 7 months" - https://markets.businessinsider.com/news/stocks/jim-simons-r...

I’m referring to RenTech’s well known Medallion fund, which I believe is now only available internally to longtime employees. Even in the article you linked, it says this fund has “continued to shine”.

If you think about it a little bit...And you read "Fooled By Randomness", there are 20 other tricks they could be playing here...Instead of "predicting" the market.

Re: TimesFM: Time Series Foundation Model for time-series forecasting

#117

Post by itsoktocry »

pvorb wrote: post #112
amelius wrote: post #107

Earlier quoted context omitted.

Maybe that would be a good thing. I wouldn't mourn the destruction of the stock market as it's just a giant wealth-gap increasing casino. Trading has nothing to do with underlying value.

I fully agree. The stock market is just a giant machine that pulls money out of systems.

>The stock market is just a giant machine that pulls money out of systems.

So you think the multi-trillion dollar stock market, consisting of thousands of global companies, has no use beyond "pulling money out of systems"? Weird.

Re: TimesFM: Time Series Foundation Model for time-series forecasting

#118

Post by amelius »

itsoktocry wrote: post #117
pvorb wrote: post #112

Earlier quoted context omitted.

I fully agree. The stock market is just a giant machine that pulls money out of systems.

> The stock market is just a giant machine that pulls money out of systems. So you think the multi-trillion dollar stock market, consisting of thousands of global companies, has no use beyond "pulling money out of systems"? Weird.

https://en.wikipedia.org/wiki/Tulip_mania

Just to say, weirdness happens.

Re: TimesFM: Time Series Foundation Model for time-series forecasting

#119

Post by icapybara »

frankc wrote: post #105
icapybara wrote: post #101

Earlier quoted context omitted.

Agreed, if stock prices were predictable by some technical means, they would be quickly driven to unpredictability by people trading on those technical indicators.

This is that old finance chestnut. Two finance professors are walking down the hall and one of them spots a twenty dollar bill. He goes to pick it up but the other professor stops him and says "no don't bother. If there was twenty dollars there someone would have already picked it up" Yes, people arbitrage away these anomalies, and make billions doing it.

My point was more that the arbitrage is not systematically predictable, not that the arbitrage doesn’t exist.

Re: TimesFM: Time Series Foundation Model for time-series forecasting

#120

Post by cma »

whimsicalism wrote: post #104
cma wrote: post #94

Earlier quoted context omitted.

Watch this talk from Albert Gu: Efficiently Modeling Long Sequences with Structured State Spaces https://www.youtube.com/watch?v=luCBXCErkCs They made one of the best time series models and it later became one of the best language models too (Mamba).

I have already watched that talk and know Albert Gu. His work is not about a “foundational” time series model but rather a task specific one.

OK right, same architecture, but different trained models.
Post reply on HN