Source URL: https://cloud.google.com/blog/products/data-analytics/bigquery-ml-timesfm-models-now-in-preview/
Source: Cloud Blog
Title: Zero-shot forecasting in BigQuery with the TimesFM foundation model
Feedly Summary: Accurate time-series forecasting is essential for many business scenarios such as planning, supply chain management, and resource allocation. BigQuery now embeds TimesFM, a state-of-the-art pre-trained model from Google Research, enabling powerful forecasting via the simple AI.FORECAST function.
Time-series analysis is used across a wide range of fields including retail, healthcare, finance, manufacturing, and the sciences. Through the use of forecasting algorithms, users can have a more thorough understanding of their data including the recognition of trends, seasonal variations, cyclical patterns, and stationarity.
BigQuery already natively supports the well-known ARIMA_PLUS and ARIMA_PLUS_XREG models for time-series analysis. More recently, with the rapid progress and success of large pre-trained LLM models, the Google Research team developed TimesFM, a foundational model specifically for the time series domain.
The Time Series foundation model
TimesFM is a forecasting model that’s pre-trained on a large time-series corpus of 400 billion real-world time-points. A big advantage of this model is its ability to perform “zero-shot” forecasting. This means that it can make accurate predictions on unseen datasets without any training. In terms of the architecture, TimesFM is built as a decoder-only transformer model, which outputs batches of contiguous time-point segments at a time. This model has been featured on the GIFT-Eval benchmark and Monash public dataset, with a variety of public benchmarks from different domains and granularities. While ARIMA_PLUS offers customizability and explainability, the TimesFM model provides high ease-of-use and delivers good generalizability across many business domains, and often beats custom trained statistical and deep learning models.
aside_block
How BigQuery supports TimesFM
The latest TimesFM 2.0 is now a native model in BigQuery. With 500 million parameters, TimesFM model inference runs directly on BigQuery infrastructure so there are no models to train, endpoints to manage, connections to set up, or quotas to adjust. TimesFM in BigQuery is also fast and scalable — you can forecast millions of univariate time series in a few minutes with a single SQL query.
Examples of the new AI.FORECAST function
To demonstrate, consider a use case that relies on the public bigquery-public-data.san_francisco_bikeshare.bikeshare_trips table. This dataset contains information about individual bicycle trips taken using the San Francisco Bay Area’s bike-share program.
Example 1: Single time series
The following query aggregates the total number of bike-share trips on each day and forecasts the number of trips for the next 10 days (the default horizon).
code_block
<ListValue: [StructValue([(‘code’, “SELECT *\r\nFROM\r\n AI.FORECAST(\r\n (\r\n SELECT TIMESTAMP_TRUNC(start_date, DAY) AS trip_date, COUNT(*) AS num_trips\r\n FROM `bigquery-public-data.san_francisco_bikeshare.bikeshare_trips`\r\n GROUP BY 1\r\n ),\r\n timestamp_col => ‘trip_date’,\r\n data_col => ‘num_trips’);"), (‘language’, ‘lang-sql’), (‘caption’, <wagtail.rich_text.RichText object at 0x3ee0fc962490>)])]>
The results look similar to:
The output includes the forecast timestamp and values through the columns forecast_timestamp and forecast_value. The confidence_level is default as 0.95. The prediction_interval_lower_bound and prediction_interval_upper_bound show the bounds for each forecasted point.
Example 2: Multiple time series
The AI.FORECAST function also lets you forecast multiple time series at a time as shown in the following example. The following query forecasts the number of bike share trips per subscriber type and per hour for the next month (approximately 720 hours), based on the previous four months of historical data.
code_block
<ListValue: [StructValue([(‘code’, "SELECT *\r\nFROM\r\n AI.FORECAST(\r\n (\r\n SELECT\r\n TIMESTAMP_TRUNC(start_date, HOUR) AS trip_hour,\r\n subscriber_type,\r\n COUNT(*) AS num_trips\r\n FROM `bigquery-public-data.san_francisco_bikeshare.bikeshare_trips`\r\n WHERE start_date >= TIMESTAMP(‘2018-01-01’)\r\n GROUP BY 1, 2\r\n ),\r\n horizon => 720,\r\n timestamp_col => ‘trip_hour’,\r\n data_col => ‘num_trips’,\r\n id_cols => [‘subscriber_type’]);"), (‘language’, ‘lang-sql’), (‘caption’, <wagtail.rich_text.RichText object at 0x3ee0fc9622b0>)])]>
The results look similar to the following:
In addition to the columns used by the single time series, there’s the time series identifier column, which we defined earlier as subscriber_type.Visualize the results
You can merge the history data and forecasted data and visualize these results. The following graph visualizes the ‘Subscriber’ time series with its the lower and upper bounds of the prediction interval as follows:
You can see the detailed queries we used to generate this in the tutorial.
When to use TimesFM vs ARIMA_PLUS
For quick, out-of-the-box forecasts, establishing baselines, or identifying general trends with minimal setup, use TimesFM. If you need to model specific patterns, fine-tune forecasts for seasonality or holidays, multivariate (ARIMA_PLUS_XREG), require explainable results, or want to leverage a longer historical context, ARIMA_PLUS is the more suitable choice.
Take the next step
The TimesFM 2.0 model is now available in BigQuery in preview. For more details, please see the tutorial and the documentation.
AI Summary and Description: Yes
**Summary:** The text discusses Google’s TimesFM, a pre-trained forecasting model now available in BigQuery that enhances time-series analysis capabilities for various business applications. It highlights the ease of use, speed, and scalability of TimesFM, allowing accurate forecasting with minimal setup compared to traditional models.
**Detailed Description:**
The content is primarily focused on the introduction of TimesFM, a new state-of-the-art forecasting model integrated into Google BigQuery, which facilitates advanced time-series analysis useful for numerous sectors. Below are the key points that illustrate its significance:
– **Importance of Time-series Forecasting:**
– Critical for various business functions including planning, supply chain management, and resource allocation.
– Applicable across diverse fields such as retail, healthcare, finance, manufacturing, and scientific research.
– **Overview of TimesFM:**
– Developed by Google Research, TimesFM is pre-trained on a vast dataset comprising 400 billion time-points.
– Capable of “zero-shot” forecasting, meaning it can generate predictions on new datasets without prior training.
– **Technical Specifications:**
– Built on a decoder-only transformer architecture with 500 million parameters.
– Inference runs directly within BigQuery infrastructure, eliminating the need for model training, endpoint management, or complex setup.
– **AI.FORECAST Function:**
– Introduces a user-friendly SQL function that simplifies the forecasting process.
– Allows for the forecasting of both single and multiple time series, demonstrating versatility through practical examples involving bike-share trip data.
– **Use Cases of TimesFM:**
– Single series example: Forecasting daily bike-share trip totals for the upcoming 10 days.
– Multiple series example: Predicting hourly bike-share trips based on subscriber type, enhancing insights into user behavior.
– **Comparison with ARIMA_PLUS:**
– TimesFM is suited for rapid forecasts with minimal configuration, ideal for baseline estimations and trend identification.
– ARIMA_PLUS provides higher customizability and explainability for more complex forecasting needs, especially involving specific patterns or seasonal effects.
– **Scalability and Performance:**
– Highlights its capability to handle millions of univariate time series forecasts efficiently, making it a scalable solution for large-scale analytical needs.
– **Conclusion:**
– TimesFM 2.0 is positioned as a crucial addition to BigQuery, providing users with advanced forecasting tools that are easy to implement and run efficiently within existing infrastructure.
This innovative model significantly enhances forecasting capabilities for businesses, making it valuable for security and compliance professionals who may want to leverage such predictive analytics for operational risk management and strategic decision-making.