About

Jason Eshraghian is an Assistant Professor at the Department of Electrical and Computer Engineering at the University of California, Santa Cruz. His research focuses on brain-inspired circuit design aimed at accelerating artificial intelligence algorithms and spiking neural networks. He leads the UCSC Neuromorphic Computing Group, which is dedicated to advancing neuromorphic engineering and computing technologies. For ongoing research updates, the lab maintains a dedicated webpage at ncg.ucsc.edu.

Research topics

• Computer Science
• Artificial Intelligence
• Engineering
• Computer architecture
• Electronic engineering
• Neuroscience
• Psychology
• Embedded system
• Cognitive science
• Computer hardware
• Electrical engineering

Selected publications

• NeuroAI and Beyond: Bridging Between Advances in Neuroscience and ArtificialIntelligence

  ArXiv.org · 2026-04-19

  articleOpen access

  Neuroscience and Artificial Intelligence (AI) have made impressive progress in recent years but remain only loosely interconnected. Based on a workshop convened by the National Science Foundation in August 2025, we identify three fundamental capability gaps in current AI: the inability to interact with the physical world, inadequate learning that produces brittle systems, and unsustainable energy and data inefficiency. We describe the neuroscience principles that address each: co-design of body and controller, prediction through interaction, multi-scale learning with neuromodulatory control, hierarchical distributed architectures, and sparse event-driven computation. We present a research roadmap organized around these principles at near, mid, and long-term horizons. We argue that realizing this program requires a new generation of researchers trained across the boundary between neuroscience and engineering, and describe the institutional conditions: interdisciplinary training, hardware access, community standards, and ethics, needed to support them. We conclude that NeuroAI, neuroscience-informed artificial intelligence, has the potential to overcome limitations of current AI while deepening our understanding of biological neural computation.

  PublisherOA PDF

• Higher-order neuromorphic Ising machines—autoencoders and Fowler-Nordheim annealers are all you need for scalability

  Nature Communications · 2026-04-16

  articleOpen access

  We report that an autoencoder-based neuromorphic architecture, combined with Fowler-Nordheim annealing, is sufficient to implement scalable higher-order Ising machines. We show that these machines can consistently produce state-of-the-art solutions with high reliability and with competitive time-to-solution metrics. The autoencoder captures higher-order interactions by decomposing Ising clauses and Ising spins into encoder-decoder layers of spiking neurons, thereby keeping the resource complexity independent of the interaction order for sparse problems. An annealing process based on the dynamics of Fowler-Nordheim quantum mechanical tunneling extrapolates between an $${{\mathcal{O}}}(1/t)$$ annealing schedule and an $${{\mathcal{O}}}(1/\log (t))$$ annealing schedule. This not only ensures fast convergence towards high-quality solutions but also guarantees asymptotic convergence to the Ising ground state. To demonstrate the advantages of the proposed higher-order neuromorphic Ising machine, we systematically solved benchmark combinatorial optimization problems such as MAX-CUT and MAX-SAT, comparing the results to those obtained using a second-order Ising machine employing the same annealing process. The authors demonstrate that an autoencoder-based neuromorphic architecture combined with Fowler-Nordheim annealing, is sufficient to implement scalable higher-order Ising machines. They show that these machines can consistently produce state-of-the-art solutions with high reliability and competitive time-to-solution metrics.

  PublisherOA PDFDOI

• Efficient knowledge distillation via salient feature masking

  APL Machine Learning · 2026-02-06

  articleOpen accessSenior author

  Traditional Knowledge Distillation (KD) transfers all outputs from a teacher model to a student model, often introducing knowledge redundancy. This redundancy dilutes critical information, leading to degraded student model performance. To address this, we propose Salient Feature Masking for Knowledge Distillation (SFKD), a lightweight enhancement that masks out less informative components and selectively distills only the top-K activations. SFKD is a drop-in modification applicable to both logit-based and feature-based KD, incurs negligible overhead, and sharpens the student’s learning signal. Empirically, SFKD yields consistent gains over strong KD baselines across architectures (ConvNeXt, ViT) and datasets (CIFAR-100: up to +2.43 pp; CUB-200: up to +6.39 pp; ImageNet-1K: up to +3.57 pp). We also provide intuition from the information bottleneck perspective to motivate why filtering out less salient teacher signals benefits the student. Overall, SFKD is a simple, empirically validated method for training student models that are both leaner and more accurate.

  PublisherDOI

• NeuroAI and Beyond: Bridging Between Advances in Neuroscience and ArtificialIntelligence

  Cold Spring Harbor Laboratory Institutional Repository (Cold Spring Harbor Laboratory) · 2026-04-19

  preprintOpen access

  Neuroscience and Artificial Intelligence (AI) have made impressive progress in recent years but remain only loosely interconnected. Based on a workshop convened by the National Science Foundation in August 2025, we identify three fundamental capability gaps in current AI: the inability to interact with the physical world, inadequate learning that produces brittle systems, and unsustainable energy and data inefficiency. We describe the neuroscience principles that address each: co-design of body and controller, prediction through interaction, multi-scale learning with neuromodulatory control, hierarchical distributed architectures, and sparse event-driven computation. We present a research roadmap organized around these principles at near, mid, and long-term horizons. We argue that realizing this program requires a new generation of researchers trained across the boundary between neuroscience and engineering, and describe the institutional conditions: interdisciplinary training, hardware access, community standards, and ethics, needed to support them. We conclude that NeuroAI, neuroscience-informed artificial intelligence, has the potential to overcome limitations of current AI while deepening our understanding of biological neural computation.

  OA PDFDOI

• Publisher’s Note: “Erratum: ‘Brain-inspired learning in artificial neural networks: A review’ [APL Mach. Learn. 2, 021501 (2024)]” [APL Mach. Learn. 3, 039901 (2025)]

  APL Machine Learning · 2026-04-13

  articleOpen accessSenior author
  PublisherDOI

• Direct Semantic Communication Between Large Language Models via Vector Translation

  ArXiv.org · 2025-11-06

  preprintOpen accessSenior author

  In multi-agent settings, such as debate, reflection, or tool-calling, large language models (LLMs) pass messages as plain tokens, discarding most latent semantics. This constrains information transfer and adds unnecessary computational overhead. We form a latent bridge via vector translations, which use learned mappings that enable direct semantic exchange between representation spaces. A dual-encoder translator trained between Llama-2-7B and Mistral-7B-Instruct attains an average cosine alignment of 0.538. Injecting the translated vectors at 30 percent blending strength steers the target model's generation without destabilizing logits. Bidirectional evaluation shows a 2.01:1 transfer asymmetry, indicating that general-purpose models yield more transferable representations than instruction-tuned variants. This conservative injection preserves computational stability while demonstrating that cross-model latent communication is feasible, enabling collaborative AI systems that share meaning rather than tokens.

  PublisherOA PDFDOI

• A predictive approach to enhance time-series forecasting

  Nature Communications · 2025-09-30 · 6 citations

  articleOpen accessSenior author

  Accurate time-series forecasting is crucial in various scientific and industrial domains, yet deep learning models often struggle to capture long-term dependencies and adapt to data distribution shifts over time. We introduce Future-Guided Learning, an approach that enhances time-series event forecasting through a dynamic feedback mechanism inspired by predictive coding. Our method involves two models: a detection model that analyzes future data to identify critical events and a forecasting model that predicts these events based on current data. When discrepancies occur between the forecasting and detection models, a more significant update is applied to the forecasting model, effectively minimizing surprise, allowing the forecasting model to dynamically adjust its parameters. We validate our approach on a variety of tasks, demonstrating a 44.8% increase in AUC-ROC for seizure prediction using EEG data, and a 23.4% reduction in MSE for forecasting in nonlinear dynamical systems (outlier excluded). By incorporating a predictive feedback mechanism, Future-Guided Learning advances how deep learning is applied to time-series forecasting.

  PublisherOA PDFDOI

• Advancing Spiking Neural Networks Towards Multiscale Spatiotemporal Interaction Learning

  Proceedings of the AAAI Conference on Artificial Intelligence · 2025-04-11 · 5 citations

  articleOpen access

  Recent advancements in neuroscience research have propelled the development of Spiking Neural Networks (SNNs), which not only have the potential to further advance neuroscience research but also serve as an energy-efficient alternative to Artificial Neural Networks (ANNs) due to their spike-driven characteristics. However, previous studies often overlooked the multiscale information and its spatiotemporal correlation between event data, leading SNN models to approximate each frame of input events as static images. We hypothesize that this oversimplification significantly contributes to the performance gap between SNNs and traditional ANNs. To address this issue, we have designed a Spiking Multiscale Attention (SMA) module that captures multiscale spatiotemporal interaction information. Furthermore, we developed a regularization method named Attention ZoneOut (AZO), which utilizes spatiotemporal attention weights to reduce the model's generalization error through pseudo-ensemble training. Our approach has achieved state-of-the-art results on mainstream neuromorphic datasets. Additionally, we have reached a performance of 77.1\% on the Imagenet-1K dataset using a 104-layer ResNet architecture enhanced with SMA and AZO. This achievement confirms the state-of-the-art performance of SNNs with non-transformer architectures and underscores the effectiveness of our method in bridging the performance gap between SNN models and traditional ANN models.

  PublisherOA PDFDOI

• A Systematic Analysis of Hybrid Linear Attention

  ArXiv.org · 2025-07-08

  preprintOpen accessSenior author

  Transformers face quadratic complexity and memory issues with long sequences, prompting the adoption of linear attention mechanisms using fixed-size hidden states. However, linear models often suffer from limited recall performance, leading to hybrid architectures that combine linear and full attention layers. Despite extensive hybrid architecture research, the choice of linear attention component has not been deeply explored. We systematically evaluate various linear attention models across generations - vector recurrences to advanced gating mechanisms - both standalone and hybridized. To enable this comprehensive analysis, we trained and open-sourced 72 models: 36 at 340M parameters (20B tokens) and 36 at 1.3B parameters (100B tokens), covering six linear attention variants across five hybridization ratios. Benchmarking on standard language modeling and recall tasks reveals that superior standalone linear models do not necessarily excel in hybrids. While language modeling remains stable across linear-to-full attention ratios, recall significantly improves with increased full attention layers, particularly below a 3:1 ratio. Our study highlights selective gating, hierarchical recurrence, and controlled forgetting as critical for effective hybrid models. We recommend architectures such as HGRN-2 or GatedDeltaNet with a linear-to-full ratio between 3:1 and 6:1 to achieve Transformer-level recall efficiently. Our models are open-sourced at https://huggingface.co/collections/m-a-p/hybrid-linear-attention-research-686c488a63d609d2f20e2b1e.

  PublisherOA PDFDOI

• Spiking neural networks on FPGA: A survey of methodologies and recent advancements

  Neural Networks · 2025-02-14 · 21 citations

  reviewOpen accessSenior author
  PublisherOA PDFDOI

Frequent coauthors

• Herbert Ho‐Ching Iu

  University of Western Australia

  74 shared

• Omid Kavehei

  University of Sydney

  56 shared

• Xiaoyuan Wang

  China Aerospace Science and Technology Corporation

  53 shared

• Armin Nikpour

  Royal Prince Alfred Hospital

  50 shared

• Nhan Duy Truong

  University of Sydney

  49 shared

• Mostafa Rahimi Azghadi

  39 shared

• Sung-Mo Kang

  37 shared

• Corey Lammie

  IBM Research - Zurich

  34 shared

Labs

• Neuromorphic Computing GroupPI

Education

• Bachelor of Laws & Electronic Engineering (Combined Degree)

  University of Western Australia

  2016

Awards & honors

• 2023 IEEE CAS Darlington Best Paper Award
• 2020 IEEE ICECS Best Live Demonstration Award
• 2019 IEEE TVLSI Best Paper Award
• 2019 IEEE AICAS Best Paper Award
• Fulbright Research Fellowship (Australian-American Fulbright…