# Adaptive Thresholding for Multi-Label Classification via Global-Local Signal Fusion

arXiv: https://arxiv.org/abs/2505.03118 · HTML: https://mlsanity.com/paper/2505.03118/

## Key facts

- Cited by 2 papers in the map (data through 14 Oct 2026).
- Builds on 3 of its 9 references that are in the map.
- First posted on arXiv 6 May 2025.

## Builds on (most cited first, 3 of 3)

- [Character-level Convolutional Networks for Text Classification](https://mlsanity.com/paper/1509.01626/) (2015)
- [AttentionXML: Label Tree-based Attention-Aware Deep Model for High-Performance Extreme Multi-Label Text Classification](https://mlsanity.com/paper/1811.01727/) (2018)
- [Long-tailed Extreme Multi-label Text Classification with Generated Pseudo Label Descriptions](https://mlsanity.com/paper/2204.00958/) (2022)

## Built on by (most cited first, 2 of 2)

- [RAPT: Retrieval-Augmented Post-hoc Thresholding for Multi-Label Classification](https://mlsanity.com/paper/2605.16535/) (2026)
- [FrugalSOT - Frugal Search Over the Models](https://mlsanity.com/paper/2608.21621/) (2026)

## Abstract (arXiv)

Multi-label classification (MLC) requires predicting multiple labels per sample, often under heavy class imbalance and noisy conditions. Traditional approaches apply fixed thresholds or treat labels independently, overlooking context and global rarity. We introduce an adaptive thresholding mechanism that fuses global (IDF-based) and local (KNN-based) signals to produce per-label, per-instance thresholds. Instead of applying these as hard cutoffs, we treat them as differentiable penalties in the loss, providing smooth supervision and better calibration. Our architecture is lightweight, interpretable, and highly modular. On the AmazonCat-13K benchmark, it achieves a macro-F1 of 0.1712, substantially outperforming tree-based and pretrained transformer-based methods. We release full code for reproducibility and future extensions.