RobustNCL: Robustifying Neuromorphic Continual Learning under Transient Faults with Lightweight Detect-and-Cure Strategy

Rachmad Vidya Wicaksana Putra; Shuochen Zhou; Muhammad Shafique

cs.NEOct 22, 2026

RobustNCL: Robustifying Neuromorphic Continual Learning under Transient Faults with Lightweight Detect-and-Cure Strategy

Authors: Rachmad Vidya Wicaksana Putra, Shuochen Zhou, Muhammad Shafique

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

Abstract

In neural networks, continual learning (CL) is a prominent approach for enabling deployed systems to adapt to dynamic operational environments by effectively learning from new unseen data at runtime. Recently, the neuromorphic continual learning (NCL) concept has emerged as an energy-efficient alternative to CL by leveraging sparse event-based operations. However, NCL performance can be undermined by transient faults in the hardware platforms, which may come from high-energy particle strikes at any time. Currently, NCL reliability under transient faults has not been studied, and hence the corresponding fault-tolerance techniques remain unexplored. To address this, we propose a novel RobustNCL framework to improve NCL reliability in the presence of transient faults in the DRAM-based off-chip memory of hardware platforms through a lightweight detect-and-cure strategy. Specifically, it first monitors the presence of faults during weight data access from off-chip memory, and then cures the accessed weights if anomalous signatures are detected. It provides three curing variants: replacement with zero, min/max value, and most-probable value from the original/clean weight distribution. To support this, the weight curing values are stored in the radiation-hardened memory. Experimental results show that, in the presence of high fault rates (e.g., 5x10^{-4} for memory replay), RobustNCL maintains high accuracy close to the accuracy of fault-free models (i.e., within -0.5% accuracy for the best RobustNCL variant) and significantly improves accuracy from the baseline without fault mitigation across different NCL methods and datasets (i.e., Split MNIST, Split Fashion-MNIST, Split CIFAR-10, and Split CIFAR-100), while incurring negligible latency and energy overheads. These results highlight that our RobustNCL framework is a promising approach for enabling reliable and energy-efficient NCL systems.

Figures & tables

Figure 1: RobustNCL effectively improves NCL reliability againts transient faults in the DRAM off-chip/main memory under a task-incremental learning (TIL) scenario across different benchmarks: (a) Split MNIST, (b) Split Fashion-MNIST, and (c) Split CIFAR-10, across different NCL methods: task-specific subnetwork (Subnet) and memory replay (MR). The fault rate is denoted as

f_{R}

considering a DDR3-1600 2Gb DRAM chip ( Micron, 2010 ) , RobustNCL variants are with v1-v3 codes.

Figure 2: The RobustNCL framework, showing the transient fault exposure stage, our detect-and-cure (D&C) strategy stage supported by weight replacement (curing) value selection, and the NCL training stage. RobustNCL variants depend on how the faulty weights are replaced by selected curing weights.

Figure 3: (a) The NCL pipeline consists of three phases: the offline training at design time, the NCL training at runtime, and the inference at runtime. (b) DRAM organization overview, including channels, ranks, chips, and banks.

Figure 4: Overview of (a) the task-specific subnetwork (Subnet) method and (b) the memory replay (MR) method that are considered in this paper.

Figure 5: Transient-fault modeling in the DRAM with several key steps: flattening multiple banks into a continuous memory space, injecting transient faults, and curing anomalous weights. Transient fault injection can happen at any time, as illustrated in Fig. 2 .

Figure 6: Experimental results from different NCL methods: (a) Subnet and (b) MR. These results show weight distributions of both the fault-free and faulty models from their layer 2 at certain model states (i.e., after training task 1 and task 2), inference tasks (i.e., task 1 and task 2), and fault rates ( $f_{R}=1{\times}10^{-5}$ for Subnet and $f_{R}=1{\times}10^{-4}$ for MR).

Dataset	Technique	$5{\times}10^{-9}$	$1{\times}10^{-8}$	$5{\times}10^{-8}$	$1{\times}10^{-7}$	$5{\times}10^{-7}$	$1{\times}10^{-6}$	$5{\times}10^{-6}$	$1{\times}10^{-5}$
Split MNIST	Baseline	99.84	99.83	90.05	76.38	51.28	50.32	51.28	51.28
	RobustNCLv1	99.81	99.78	99.79	99.66	99.72	99.77	99.77	99.79
	RobustNCLv2	99.81	99.78	99.80	99.66	99.74	99.79	99.80	99.82
	RobustNCLv3	99.81	99.78	99.79	99.66	99.76	99.76	99.72	99.80
Split FMNIST	Baseline	99.01	89.36	66.79	57.54	50.00	50.00	50.00	50.00
	RobustNCLv1	99.01	99.05	99.03	99.00	98.98	99.06	99.05	99.01

Table 1: Final average accuracy (%) obtained by the Subnet method on the 5-task datasets evaluated across all trained tasks (task 1 – task 5) and across different fault-mitigation techniques.

Dataset	Technique	$1{\times}10^{-7}$	$5{\times}10^{-7}$	$1{\times}10^{-6}$	$5{\times}10^{-6}$	$1{\times}10^{-5}$	$5{\times}10^{-5}$	$1{\times}10^{-4}$	$5{\times}10^{-4}$
Split MNIST	Baseline	99.40	99.37	99.38	99.47	98.61	51.50	51.50	51.50
	RobustNCLv1	99.33	99.31	99.41	99.22	99.40	99.41	99.37	99.34
	RobustNCLv2	99.39	99.35	99.19	99.44	99.49	99.30	99.28	99.51
	RobustNCLv3	99.29	99.32	99.49	99.38	99.40	99.23	99.39	99.37
Split FMNIST	Baseline	99.03	98.99	99.08	98.89	87.37	94.14	50.00	50.00
	RobustNCLv1	99.01	98.97	99.07	98.96	99.00	99.06	99.01	98.31

Table 2: Final average accuracy (%) obtained by the MR method on the 5-task datasets evaluated across all trained tasks (task 1 – task 5) and across different fault-mitigation techniques.

Technique	Subnet				MR
Technique	$1{\times}10^{-8}$	$1{\times}10^{-7}$	$1{\times}10^{-6}$	$1{\times}10^{-5}$	$1{\times}10^{-8}$	$1{\times}10^{-7}$	$1{\times}10^{-6}$	$1{\times}10^{-5}$
Baseline	60.00	10.00	10.71	10.00	66.66	66.62	10.00	10.00
RobustNCLv1	70.73	70.61	70.61	70.82	67.11	66.91	66.84	66.35
RobustNCLv2	70.73	70.66	70.50	70.59	66.75	66.35	66.59	65.90
RobustNCLv3	70.73	70.57	70.66	71.02	66.65	66.63	66.25	66.82

Table 3: Final average accuracy (%) obtained by both the Subnet and MR methods on the 10-task dataset (Split CIFAR-100) evaluated across all trained tasks (task 1 – task 10).

Neuromorphic Computing

Neuromorphic computing uses brain-inspired, event-driven algorithms and hardware to process information with low power and latency.

See all 194 papers

ENCL: Enabling Efficient Neuromorphic Continual Learning with Local Learning Rule and Sub-Network Freezing Strategy
Oct 22, 2026Rachmad Vidya Wicaksana Putra, Shuochen Zhou, Hassan Raza +1Energy-Efficient ML Local Learning
eBRAIN Lab, New York University (NYU) Abu Dhabi, United Arab Emirates
IDER: IDempotent Experience Replay for Reliable Continual Learning
Feb 28, 2026Zhanwang Liu, Yuting Li, Haoyuan Gao +4Continual Learning Catastrophic Forgetting
School of Computer Science, Shanghai Jiao Tong University · State Key Laboratory of General Artificial Intelligence, BIGAI · Lehigh University +1
Leveraging ECRAM for Edge Continual Learning
Jul 22, 2026Nabila Tasnim, Haoran Liu, Qing Cao +1Compute-in-Memory Continual Learning
University of Illinois Urbana-Champaign

Oct 22, 2026J/K move · Enter open · S save

ENCL: Enabling Efficient Neuromorphic Continual Learning with Local Learning Rule and Sub-Network Freezing Strategy

Rachmad Vidya Wicaksana Putra, Shuochen Zhou, Hassan Raza, Muhammad Shafique

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

Energy-Efficient ML Local Learning Task-Incremental Learning

Open paper AskPDF arXiv

Continual learning (CL) in neural networks is a promising learning approach for updating the knowledge of deployed systems based on new data encountered at run time, that have not been learned during the offline training phase. To enable their power/energy-efficient CL processing, neuromorphic continual learning (NCL) concept with sparse event-based operations have recently been introduced. However, state-of-the-art NCL methods typically employ the memory-hungry data/experience replay, coupled with costly global learning rule (i.e., backpropagation-based surrogate gradient). To address these limitations, we propose ENCL, a novel methodology that enables efficient NCL by employing a lightweight local learning rule and sub-network freezing strategy. ENCL leverages the temporally- and spatially-local learning rule to perform efficient credit assignment for training new tasks, alleviating the expensive cost of global learning rule. ENCL also employs a sub-network freezing strategy based on the given tasks to avoid the needs for network architecture expansion, hence further improving the efficiency and scalability of the NCL training process. To efficiently support this, we train each task with a small number of epochs as well as adjusted learning rate and timestep. Experiment results on task-incremental learning (TIL) scenario show that, our ENCL consistently achieves strong performance across different datasets as compared to state-of-the-art works. Specifically, ENCL achieves competitive accuracy comparable to that of the state-of-the-art data/experience replay, while reducing memory cost by up to 88%, NCL training time by up to 51%, average power consumption by up to 40%, and energy consumption by up to 65% across different benchmarking datasets (i.e., Split MNIST, Split Fashion MNIST, Split CIFAR-10, and Split CIFAR-100).

Abstract

Neuromorphic Computing

Explore similar work

ENCL: Enabling Efficient Neuromorphic Continual Learning with Local Learning Rule and Sub-Network Freezing Strategy

IDER: IDempotent Experience Replay for Reliable Continual Learning

Leveraging ECRAM for Edge Continual Learning