arXiv:2611.17600v1 Announce Type: new
Abstract: Multimodal tabular learning involves classification and regression on tables combining structured numerical and categorical features with unstructured text and images. These unstructured fields are commonly represented by pretrained encoder embeddings reduced with principal component analysis (PCA). This reduction prioritizes input variance but may overlook information relevant to predicting the target. We introduce LightTAR, a simple, inexpensive method that adapts dimensionality reduction to the prediction task. LightTAR augments the standard PCA representation with a few target-aware features extracted jointly from the final and an intermediate encoder layer using regularized partial least squares (PLS). We evaluate LightTAR with 21 tabular models, including tree ensembles, neural networks, and tabular foundation models. LightTAR improves average predictive performance for all models and, averaged across models, improves prediction on almost every MulTaBench dataset. These gains come with a small increase in total runtime relative to PCA representations. LightTAR integrates directly into existing encoding pipelines, improving tabular learners’ predictive performance without modifying the encoder or the learner.
Trending
- How Secure Is Your Password? Here’s How To Test It
- These iOS Settings Will Help You Control Your Mobile Data Use
- What The Tiny Key Icon Means On Your Android Device
- You Don’t Need A Foldable Phone To Use Split Screen On Android
- The Advantages Of Using A USB-C Docking Station
- Want More Control Over Your YouTube Homepage? Try Using Custom Feeds
- How Long Is A New Fitbit Actually Supposed To Last?
- Here’s When A Raspberry Pi 5 Could Be Considered Overkill
