MTLiquid: Enabling Efficient Multi-Task Learning using Liquid Neural Networks for Lightweight Healthcare Monitoring Systems

Rachmad Vidya Wicaksana Putra; Fahad Abdul Rauf; Muhammad Shafique

cs.LGSep 27, 2026

MTLiquid: Enabling Efficient Multi-Task Learning using Liquid Neural Networks for Lightweight Healthcare Monitoring Systems

Authors: Rachmad Vidya Wicaksana Putra, Fahad Abdul Rauf, Muhammad Shafique

Organizations: eBRAIN Lab, New York University (NYU) Abu Dhabi, Abu Dhabi, United Arab Emirates

Abstract

Continuous-time sensing and monitoring with timely and accurate decision-making are critical for many real-world applications. In healthcare monitoring systems, physiological signals are often available or sampled at irregular time intervals, hence requiring continuous-time processing to provide accurate prediction. Moreover, such systems often need to solve multiple detection/prediction tasks to provide a comprehensive patient review from different physiological aspects for more accurate decision-making. To solve this, continuous-time neural networks (CTNNs) can be employed. However, state-of-the-art works typically solve only one task at each network, thereby limiting their efficiency gains. To address this limitation, we propose MTLiquid, a novel methodology to enable efficient multi-task learning in continuous-time processing for healthcare monitoring systems through effective network design and training strategy. MTLiquid employs: (1) multiple input and output heads to accommodate different tasks, while sharing the same backbone network across tasks; as well as (2) an effective training strategy that leverages a loss-weighting technique to balance learning updates across different tasks and a proportional data presentation technique to address imbalanced dataset sizes. Experimental results for mortality prediction (P12) and sepsis early detection (P19) tasks for ICU patients show that, MTLiquid achieves strong performance (AUROC: 0.84 for P12 and 0.94 for P19) comparable to the state-of-the-art single-task learning in both continuous-time networks (AUROC: 0.84 for P12 and 0.95 for P19) and discrete-time networks (AUROC: 0.79-0.82 for P12 and 0.92-0.94 for P19), while incurring significantly smaller memory cost by 44%-94%. These results highlight the potential of our MTLiquid methodology to enable lightweight continuous-time healthcare monitoring systems for better decision-making.

Figures & tables

Figure 1: Multi-task architecture in MTLiquid based on the CfC concept. Task-specific inputs are projected to a common 64-dimensional space, concatenated with

h(t-1)

, and passed through a shared backbone. Candidate and time-gate branches (gated by

\sigma

over

\Delta t_{t}

) update the hidden state, which feeds task-specific linear heads.

Setting	Model	Task	Params	Size (MB)	AUROC	AUPRC
Single-task (one model per task)	CfC	P12	92,930	0.35	0.8409 $\pm$ 0.0032	0.5238 $\pm$ 0.0108
	CfC	P19	95,362	0.36	0.9472 $\pm$ 0.0083	0.7532 $\pm$ 0.0039
	LSTM	P12	875,010	3.34	0.8234 $\pm$ 0.0163	0.4624 $\pm$ 0.0299
	LSTM	P19	913,922	3.49	0.9226 $\pm$ 0.0008	0.7460 $\pm$ 0.0083
	RNN	P12	219,138	0.84	0.7882 $\pm$ 0.0080	0.4608 $\pm$ 0.0070
	RNN	P19	228,866	0.87	0.9381 $\pm$ 0.0075	0.7358 $\pm$ 0.0386

Table 1: Single-task versus multi-task performance across architectures. For multi-task, each row is the same shared model (one architecture, one parameter count). CfC matches or exceeds LSTM/RNN at a fraction of the parameters and model size.

Configuration	P12 AUROC	P12 AUPRC	P19 AUROC	P19 AUPRC
Baseline (no loss weighting, zero-fill, default cycle)	0.7175	0.2974	0.9140	0.7193
+ Loss weighting (Kendall et al.)	0.7328	0.3162	0.9322	0.7281
+ GRU-D forward-fill imputation	0.8010	0.4324	0.9348	0.7547
Loader ablation: --truncate_to_shorter	0.8415	0.5027	0.8799	0.6987
Capped cycling, $R=3$ (our MTLiquid)	0.8418	0.5132	0.9379	0.7662

Table 2: Ablation progression for the CfC multi-task backbone. Each stage is cumulative over the previous one unless noted.

Explore similar work

Jun 10, 2026

Multi-Rate Mixture of Experts for Accelerating Liquid Neural Network Training

Multivariate time-series data often exhibit complex temporal dependencies, irregular sampling, and heterogeneous dynamics across multiple time scales, making accurate sequence modeling particularly challenging. Traditional recurrent neural networks (RNNs), such as Long Short-Term Memory (LSTM) networks, operate in discrete time and may struggle to effectively capture continuous and irregular temporal behaviors. Liquid Neural Networks (LNNs) address some of these limitations through continuous-time dynamics, but standard LNN architectures typically rely on a single dynamical system, limiting their ability to model heterogeneous temporal patterns. To address these challenges, we propose a Multi-Rate Mixture-of-Experts (MR-MoE) framework built on top of Liquid Neural Networks. In the proposed architecture, multiple LNN-based experts operate at distinct time scales, enabling the model to explicitly separate fast-changing dynamics from slow-evolving temporal trends. A gating network further enables adaptive expert specialization based on input conditions. In addition, we incorporate both feature-level and temporal attention mechanisms to improve robustness, interpretability, and long-range dependency modeling. Feature-level attention suppresses noisy or irrelevant variables, while temporal attention selectively focuses on informative historical states. We evaluate the proposed framework on a complex multivariate time-series prediction task and compare it against strong baselines, including LSTM, monolithic LNN, and standard MoE models. Experimental results demonstrate that the proposed MR-MoE framework consistently achieves improved AUROC and AUPRC performance while maintaining favorable computational efficiency. These results highlight the effectiveness of combining continuous-time dynamics, multi-scale expert decomposition, and adaptive attention mechanisms for time-series modeling.

Shilong Zong, Almuatazbellah Boker, Hoda Eldardiry

Department of Computer Science, Virginia Tech Blacksburg, VA 24061 USA · Department of Electrical and Computer Engineering, Virginia Tech Blacksburg, VA 24061 USA

Irregular Time-Series Modeling Liquid Neural Networks Mixture-of-Experts Models Multiscale Time-Series Modeling

Ask PDF

Apr 12, 2026

Sense Less, Infer More: Agentic Multimodal Transformers for Edge Medical Intelligence

Edge-based multimodal medical monitoring requires models that balance diagnostic accuracy with severe energy constraints. Continuous acquisition of ECG, PPG, EMG, and IMU streams rapidly drains wearable batteries, often limiting operation to under 10 hours, while existing systems overlook the high temporal redundancy present in physiological signals. We introduce Adaptive Multimodal Intelligence (AMI), an end-to-end framework that jointly learns when to sense and how to infer. AMI integrates three components: (1) a lightweight Agentic Modality Controller that uses differentiable Gumbel-Sigmoid gating to dynamically select active sensors based on model confidence and task relevance; (2) a Learned Sigma-Delta Sensing module that applies patch-wise Delta-Sigma operations with learnable thresholds to skip temporally redundant samples; and (3) a Foundation-backed Multimodal Prediction Model built on unimodal foundation encoders and a cross-modal transformer with temporal context, enabling robust fusion even under gated or missing inputs. These components are trained jointly via a multi-objective loss combining classification accuracy, sparsity regularization, cross-modal alignment, and predictive coding. AMI is hardware-aware, supporting dynamic computation graphs and masked operations, leading to real energy and latency savings. Across MHEALTH, HMC Sleep, and WESAD datasets, it reduces sensor usage by 48.8% while improving state-of-the-art accuracy by 1.9% on average.

Chengwei Zhou, Zhaoyan Jia, Haotian Yu +6

Case Western Reserve University, USA · University of Southern California, USA

Efficient Multimodal Inference Wearable Health Monitoring Edge ML Adaptive Sampling Multimodal Sensor Fusion

Ask PDF

May 26, 2026

Comparative Analysis of Liquid Neural Networks and LSTM for Sequential Pattern Recognition: Robustness, Efficiency, and Clinical Utility

Traditional Recurrent Neural Networks (RNNs) and Long Short-Term Memory (LSTM) units operate on discrete time steps, often failing to capture the fluid temporal dynamics of real-world physical processes. Liquid Neural Networks (LNNs), specifically Closed-form Continuous-time (CfC) networks, address this by modeling the hidden state evolution as a continuous differential equation. In this paper, we conduct a comprehensive benchmarking study across four distinct sequential modalities: neuromorphic event-based data (N-MNIST), stroke-based drawing (QuickDraw), visual handwriting (IAM), and physiological time-series (PhysioNet Sepsis-3). Furthermore, we perform a rigorous stress test using temporal dropout to evaluate model robustness against missing data. Our findings reveal that LNNs consistently provide superior parameter efficiency and significantly higher robustness in natively temporal domains and clinical environments where data sparsity is prevalent. This extended preprint provides additional background on related datasets and the LNN theoretical lineage, supplemented with a detailed appendix documenting our full implementation and experimental settings.

Ye Kyaw Thu, Thazin Myint Oo, Thepchai Supnithi

National Electronics and Computer Technology Center (NECTEC), Pathumthani, Thailand · Language Understanding Lab., Yangon, Myanmar

Efficient Neural Network Inference Liquid Neural Networks Long Short-Term Memory Networks Sequence Modeling

Ask PDF