# When Smaller Wins: Dual-Stage Distillation and Pareto-Guided Compression of Liquid Neural Networks for Edge Battery Prognostics

arXiv: https://arxiv.org/abs/2601.06227 · HTML: https://mlsanity.com/paper/2601.06227/

## Key facts

- Cited by 2 papers in the map (data through 14 Oct 2026).
- Builds on 1 of its 8 references that are in the map.
- First posted on arXiv 9 Jan 2026.

## Builds on (most cited first, 1 of 1)

- [EntroLnn: Entropy-Guided Liquid Neural Networks for Operando Refinement of Battery Capacity Fade Trajectories](https://mlsanity.com/paper/2601.06195/) (2026)

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

- [TIDE: Trustworthy and Interpretable Battery Degradation Estimation with Contextual Learning and Symbolic Distillation](https://mlsanity.com/paper/2607.14640/) (2026)
- [PLAN: Parallel Liquid-Inspired Approximation Network for Efficient Representation Learning in Flexible Job Shop Scheduling](https://mlsanity.com/paper/2608.03041/) (2026)

## Abstract (arXiv)

Battery management systems increasingly require accurate battery health prognostics under strict on-device constraints. This paper presents DLNet, a practical framework with dual-stage distillation of liquid neural networks that turns a high-capacity model into compact and edge-deployable models for battery health prediction. DLNet first applies Euler discretization to reformulate liquid dynamics for embedded compatibility. It then performs dual-stage knowledge distillation to transfer the teacher model's temporal behavior and recover it after further compression. Pareto-guided selection under joint error-cost objectives retains student models that balance accuracy and efficiency. We evaluate DLNet on a widely used dataset and validate real-device feasibility on an Arduino Nano 33 BLE Sense using int8 deployment. The final deployed student achieves a low error of 0.0066 when predicting battery health over the next 100 cycles, which is 15.4% lower than the teacher model. It reduces the model size from 616 kB to 94 kB with 84.7% reduction and takes 21 ms per inference on the device. These results support a practical smaller wins observation that a small model can match or exceed a large teacher for edge-based prognostics with proper supervision and selection. Beyond batteries, the DLNet framework can extend to other industrial analytics tasks with strict hardware constraints.