In life, timing is critical. That’s especially true in business, where every organization must forecast sales, demand, revenue, and capacity requirements. Accurate and reliable time-dependent forecasts could help every organization save—and earn—billions of dollars.
Table of Contents
- How can time series AI models predict the future?
- Breaking barriers in time series forecasting
- The next wave of innovation
Time series forecasting is the bread and butter of what drives an enterprise. It involves predicting future values based on past observations collected in constant time intervals, whether daily, monthly, quarterly, or yearly. Artificial intelligence is expected to accelerate and tighten business planning with new, faster, and smaller foundation models designed for multivariable time series forecasting. These models do not need to be the equivalent of an AI sledgehammer to drive results. Small time series-based or other small foundation models trained on high-quality, curated data are more energy-efficient and can achieve the same results or better.
How can time series AI models predict the future?
Time series models can be built from scratch or adapted from existing pre-trained models and are best used for predicting outcomes in time-series data. Traditionally, AI’s large language models compute relationships among words to identify patterns in the data that can be projected forward to make better decisions.
Time series foundation models look for patterns in historical observations to “understand” a temporal process. These abstract representations are what allow the models to solve predictive tasks. The longer the time series, the better the forecast. However, these kinds of measurements pose complications in ways that words, code, and pixels do not.
Some unique challenges include:
- Continuous data: Time series data often comes continuously, such as video streaming from a self-driving car or heart rate data from a smartwatch.
- Complex correlations: Different sets of time series data are often highly correlated, complicating predictions when multiple factors interact.
- Data variety: Time series data can range from stock prices to brain waves, and compressing disparate observations into an abstract representation is an enormous challenge.
This increasing complexity means that understanding how various factors influence one another over time is crucial. For example, predicting home heater sales may require considering weather anomalies or economic indicators. The more interacting variables in a time series data set, the harder it becomes to isolate the signal that indicates future behavior.
Breaking barriers in time series forecasting
AI foundation models designed for time series forecasting can be difficult to build. The sheer scale and complexity of multi-channel data sources coupled with external variables pose significant architectural challenges for the resulting model. Non-trivial computational demands also make it challenging to train and update models accurately and within desired forecasting windows.
Notably, many existing foundational models struggle to pick up on trends revealed by rapidly evolving data patterns—a process known as temporal adaptation. Time series foundation models like MOIRAI, TimesFM, and Chronos are built on hundreds of millions of parameters, demanding significant computational resources and runtime.
The next wave of innovation
Researchers and practitioners are diligently working on new approaches to overcome these obstacles and unlock the full potential of using AI in time series forecasting. An intriguing question arises: Can smaller models pre-trained solely on limited public diverse time series datasets deliver better forecasting accuracy? The answer appears to be affirmative!
Current experiments focus on the development of “tiny” foundation models significantly under 1B parameters. These smaller models—ranging from 1M to 3M parameters—can offer substantial computational efficiency while still achieving state-of-the-art results in zero/few-shot forecasting. They also have the versatility to support cross-channel and external variables, which is critical since existing popular methods often lack this capability.
Some applications of these fast and tiny general pre-trained AI models include:
- Electricity consumption: Quick forecasting of demand patterns.
- Anomaly detection: Detecting deviations from normal behavior by training on datasets featuring both regular patterns and anomalies.
Increasingly, these small models, when combined with enterprise data, can yield significant results, providing task-specific performance that rivals larger models at a fraction of the cost. They are poised to become the ‘workhorses’ of enterprise AI.
Looking ahead, AI is anticipated to facilitate a radical transformation in the business landscape. Although most of the world’s public data currently feeds existing models, a vast majority of enterprise data remains untapped. Small, fast foundation models—which have flexibility, low development costs, and wide-ranging applications—are set to play a pivotal role in this transition.
We feature the best AI tools currently available.
This article was produced as part of TechRadarPro’s Expert Insights channel where we feature the best and brightest minds in the technology industry today. The views expressed here are those of the author and are not necessarily those of TechRadarPro or Future plc. If you are interested in contributing find out more here: https://www.techradar.com/news/submit-your-story-to-techradar-pro
Leave a comment