Andrés Rojas Pastor

Consultor por vocación, intento jugar golf en ocasiones sin buenos resultados

[Marcos Galperin @marcos_galperin]

@Mercadolibre is the only public company in the world (out of +83,000 public companies) to grow more than 22 consecutive quarters in a row at a yearly rate greater than 30%. Currently 27 consecutive quarters.

@Mercadolibre es la única compañía en el mundo (de 83 mil empresas cotizantes) que crece 22 trimestres consecutivos por encima del 30% anual. Actualmente 27 trimestres consecutivos.

[Zoe Wang @zoewangai]

The results show that the ablation methods either match or outperform the original LLM-based methods in the majority of cases, across both the MAE and MSE metrics. This holds true even when the LLM parameters are randomly initialized, suggesting that the language model pretraining does not provide significant benefits for time series forecasting.

full paper: https://t.co/zFaPkre7PF

[Valeriy M., PhD, MBA, CQF @predict_addict]

The paper we have been waiting for essentially shows that
#timeseries #llms do not work in forecasting.

Back in 2022, paper “Are Transformers Effective for Time Series Forecasting?“ challenged the appearing narrative that transformers are useful for forecasting. By removing transformer elements the authors showed the performance went up ⬆️

And now people did the same with time series LLMs. The papers demonstrated:

- removing the LLM component or replacing it with a basic attention layer does not degrade the forecasting results—in most cases the results even improved!

- in fact removing even removing the language model entirely, yields comparable or better performance!

- these simpler methods after removal of LLM component reduce training and inference time by up to three orders of magnitude while maintaining comparable performance!

- the sequence modeling capabilities of LLMs do not transfer to time series. By shuffling input time series the authors find no appreciable change in performance.

What this says is that LLMs can’t deal with critical features of time series, the time order is key and if LLMs performance doesn’t change when shuffling data it basically means it doesn’t model time series.

These finding are as damming to time series LLMs as the “Are Transformers Effective for Time Series Forecasting?” was for transformers.

#timeseries
#forecasting

[Andrés Rojas Pastor @AndresRojasP]

@SomosGlobal66 @victorc64631637 @SomosGlobal66 Seria honesto de su parte indicarnos si la funcionalidad de cargar dinero en MXN la eliminaron

[Andrés Rojas Pastor @AndresRojasP]

@LaHJCK La 29 Hammerklavier la sumaría al listado.

[Marcos Galperin @marcos_galperin]

Siempre interesante escuchar la visión de B Evans sobre Inteligencia Artificial y tech en general https://t.co/UVt0NfZE3Y

[El Barroquista @elbarroquista]

Si te ha interesado la historia, se agradece si la compartes y la haces llegar a más gente.

Un RETUIT o un LIKE en este primer tuit ayudan a que llegue más.

Gracias por leerme.

https://t.co/M4Chg58DJz

[Kirk Borne @KirkDBorne]

Create and deploy #AI agents with @AbacusAI at https://t.co/X3pEy4w9mb — agents can chain user code, data transformations, models, and prompts. Learn how in this tutorial: https://t.co/EU6RXLeWuO
————
#DeepLearning #DataScience #MachineLearning #GenerativeAI #DataScientists #LLMs

[Vala Afshar @ValaAfshar]

Scientists in Japan found an enzyme that quickly eats plastic waste - it takes 400 years for a plastic bottle to decompose

[Daimler Truck @DaimlerTruck]

For Level 3 automated driving a lot of technical requirements are needed, like LiDAR, differential GPS and a multi-purpose camera. Learn more about our innovative system and find out how it works. 👇

[Mike Tamir, PhD @MikeTamir]

“examples of popular machine learning algorithms implemented in Python with mathematics behind them being explained” https://t.co/gG76ElTLVq
#AI #DeepLearning #MachineLearning #DataScience

[Andrés Rojas Pastor @AndresRojasP]

@hyperconectado @danifmartinez96 Deberian los periodistas preguntarle a @danifmartinez96 qué le decía a Egan?

[Andrés Rojas Pastor @AndresRojasP]

@aramh4ck Comparto, yo tengo formación en Administración, me enrole con el tema de Machine Learning y hoy soy manager de ML en una empresa regional. Se necesita querer aprender y disciplina.

[Elon Musk @elonmusk]

@utsavtechie Prototypes are easy, production is hard

[Vlad Haltakov @haltakov]

How Does a Self-Driving Car Work? 🔧 🧠 🚙

This is classical self-driving car software stack. Nowadays, all steps in the pipeline are dominated by machine learning.

Read below for details on each step 👇

[Andrés Rojas Pastor @AndresRojasP]

@naxxxos @AgustinLaje @GloriaAlvarez85 Gracias, el debate acabo desde que planteó los dos silogismos.

[Andrés Rojas Pastor @AndresRojasP]

@naxxxos @AgustinLaje @GloriaAlvarez85 Dónde lo puedo ver?

[Andrés Rojas Pastor @AndresRojasP]

@MJVillavecesN @fceunal @UNALOficial @ParradoYesenia Presencial? O virtual y se puede asistir a escuchar?

[Andrés Rojas Pastor @AndresRojasP]

@aramh4ck Andrés este man habla de que no es para todos y que de alguna manera puede ser bueno si se cumplen condiciones. https://t.co/lLdN39T89p qué opina?

[Andrés Rojas Pastor @AndresRojasP]

@anapastorjulian Gracias, igualmente.