About

Mohsen Imani is an Associate Professor in the Department of Computer Science at UC Irvine and serves as a director of Bio-Inspired Architecture and Systems (BIASLab). His research group focuses on practical problems in bio-inspired computing, machine learning, computer architecture, and embedded systems. Their goal is to design real-time, robust, and secure cognitive computing systems. He holds a Ph.D. in Computer Science and Engineering from UC San Diego. Mohsen Imani has been recognized with awards such as the 2024 Young Investigator Award and the DARPA Young Faculty Award, reflecting his contributions to the field.

Research topics

• Artificial Intelligence
• Computer Science
• Machine Learning
• Theoretical computer science
• Algorithm
• Database
• Mathematics
• Parallel computing

Selected publications

• SenseHD: A 40 μ <i>W</i> 8-bit Accelerator Wake-Up Circuit for Always-on Smart IoT Sensor Monitoring

  IEEE Transactions on Computer-Aided Design of Integrated Circuits and Systems · 2026-01-01

  articleSenior author

  IoT sensors and devices heavily rely on cloud-based processing, especially in healthcare monitoring, for sensor data analysis, which incurs high data transfer costs and energy consumption. To reduce constant data transmission, always-on smart sensors can use lightweight brain-inspired Hyperdimensional computing (HDC) models to efficiently classify sensor data that require further resource-intensive cloud processing. The proposed architecture maintains HDC algorithm accuracy with 8-bit integer precision for complex datasets, where 1-bit quantization is not feasible. We propose SenseHD, an all-digital CMOS always-on smart sensor accelerator for energy-constrained smart sensing applications. SenseHD achieves an extremely small power consumption of 39.21 - 92.59 μ<italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">W</i> depending on design parameterization, supporting 8-bit hypervectors and an energy-efficient near-memory fused item memory + multiply-accumulate systolic array encoder. We introduce a novel systolic similarity search unit that utilizes a small systolic array to calculate hypervector cosine similarity metrics, consuming only 110.89 nJ/inference. By intelligently transmitting only relevant data, SenseHD reduces system-level energy consumption by up to 8.8× compared to a cloud-centric baseline. Notably, the accelerator overhead is negligible, consuming only 1.2% of the total system energy.

  PublisherDOI

• Generalized Holographic Reduced Representations

  IEEE Transactions on Artificial Intelligence · 2026-01-01 · 1 citations

  articleOpen accessSenior author

  Hyperdimensional Computing (HDC) is a computationally and data-efficient paradigm that acts as a bridge between connectionist and symbolic approaches to artificial intelligence (AI). However, HDC’s simplicity poses challenges for encoding complex compositional structures, especially in its binding operation. To address this, we propose Generalized Holographic Reduced Representations (GHRR), an extension of Fourier Holographic Reduced Representations (FHRR), a specific HDC implementation. GHRR introduces a flexible, noncommutative binding operation, enabling improved encoding of complex data structures while preserving HDC’s desirable properties of robustness and transparency. In this work, we introduce the GHRR framework, prove its theoretical properties and its adherence to HDC properties, explore its kernel and binding characteristics, and perform empirical experiments showcasing its flexible non-commutativity, enhanced decoding accuracy for compositional structures. We also demonstrate that binding in GHRR is more expressive than that in other HDC variants; in particular, we show that binding in GHRR can implement a kind of attention mechanism. We verify this by replacing the attention mechanism in a transformer with its GHRR-equivalent and testing it on a language modeling task, showing improved performance compared to a vanilla transformer.

  PublisherDOI

• State-Centric Decision Process

  ArXiv.org · 2026-05-12

  articleOpen accessSenior author

  Language environments such as web browsers, code terminals, and interactive simulations emit raw text rather than states, and provide none of the runtime structure that MDP analysis requires. No explicit state space, no observation-to-state mapping, no certified transitions, and no termination criterion. We introduce the State-Centric Decision Process (SDP), a runtime framework that constructs these missing inputs by having the agent build them, predicate by predicate, as it acts. At each step the agent commits to a natural-language predicate describing how the world should look, takes an action to make it true, and checks the observation against it. Predicates that pass become certified states, and the resulting trajectory carries the four objects language environments do not provide, namely a task-induced state space, an observation-to-state mapping, certified transitions, and a termination criterion. We evaluate SDP on five benchmarks spanning planning, scientific exploration, web reasoning, and multi-hop question answering. SDP achieves the best training-free results on all five, with the advantage widening as the horizon grows. The certified trajectories additionally support analyses unavailable to reactive agents, including per-predicate credit assignment, failure localization, partial-progress measurement, and modular operator replacement.

  PublisherOA PDF

• $n$-Musketeers: Reinforcement Learning Shapes Collaboration Among Language Models

  Open MIND · 2026-02-09

  preprintSenior author

  Recent progress in reinforcement learning with verifiable rewards (RLVR) shows that small, specialized language models (SLMs) can exhibit structured reasoning without relying on large monolithic LLMs. We introduce soft hidden-state collaboration, where multiple heterogeneous frozen SLM experts are integrated through their internal representations via a trainable attention interface. Experiments on Reasoning Gym and GSM8K show that this latent integration is competitive with strong single-model RLVR baselines. Ablations further reveal a dual mechanism of expert utilization: for simpler arithmetic domains, performance gains can largely be explained by static expert preferences, whereas more challenging settings induce increasingly concentrated and structured expert attention over training, indicating emergent specialization in how the router connects to relevant experts. Overall, hidden-state collaboration provides a compact mechanism for leveraging frozen experts, while offering an observational window into expert utilization patterns and their evolution under RLVR.

  DOI

• Tell Me What to Track: Infusing Robust Language Guidance for Enhanced Referring Multi-Object Tracking

  2026-04-21

  preprintOpen access

  Referring multi-object tracking (RMOT) is an emerging cross-modal task that aims to localize an arbitrary number of targets based on a language expression and continuously track them in a video. This intricate task involves reasoning on multi-modal data and precise target localization with temporal association. However, prior studies overlook the imbalanced data distribution between newborn targets and existing targets due to the nature of the task. In addition, they only indirectly fuse multi-modal features, struggling to deliver clear guidance on newborn target detection. To solve the above issues, we conduct a collaborative matching strategy to alleviate the impact of the imbalance, boosting the ability to detect newborn targets while maintaining tracking performance. In the encoder, we integrate and enhance the cross-modal and multi-scale fusion, overcoming the bottlenecks in previous work, where limited multi-modal information is shared and interacted between feature maps. In the decoder, we also develop a referring-infused adaptation that provides explicit referring guidance through the query tokens. The experiments showcase the superior performance of our model (+3.42%) compared to prior works, demonstrating the effectiveness of our designs.

  PublisherOA PDFDOI

• Approximate quantum search

  Physics Letters A · 2026-04-06

  articleOpen accessSenior author
  PublisherDOI

• $n$-Musketeers: Reinforcement Learning Shapes Collaboration Among Language Models

  arXiv (Cornell University) · 2026-02-09

  articleOpen accessSenior author

  Recent progress in reinforcement learning with verifiable rewards (RLVR) shows that small, specialized language models (SLMs) can exhibit structured reasoning without relying on large monolithic LLMs. We introduce soft hidden-state collaboration, where multiple heterogeneous frozen SLM experts are integrated through their internal representations via a trainable attention interface. Experiments on Reasoning Gym and GSM8K show that this latent integration is competitive with strong single-model RLVR baselines. Ablations further reveal a dual mechanism of expert utilization: for simpler arithmetic domains, performance gains can largely be explained by static expert preferences, whereas more challenging settings induce increasingly concentrated and structured expert attention over training, indicating emergent specialization in how the router connects to relevant experts. Overall, hidden-state collaboration provides a compact mechanism for leveraging frozen experts, while offering an observational window into expert utilization patterns and their evolution under RLVR.

  PublisherOA PDF

• Fair Context Learning for Evidence-Balanced Test-Time Adaptation in Vision-Language Models

  ArXiv.org · 2026-02-02

  articleOpen accessSenior author

  Vision-Language Models (VLMs) such as CLIP enable strong zero-shot recognition but suffer substantial degradation under distribution shifts. Test-Time Adaptation (TTA) aims to improve robustness using only unlabeled test samples, yet most prompt-based TTA methods rely on entropy minimization -- an approach that can amplify spurious correlations and induce overconfident errors when classes share visual features. We propose Fair Context Learning (FCL), an episodic TTA framework that avoids entropy minimization by explicitly addressing shared-evidence bias. Motivated by our additive evidence decomposition assumption, FCL decouples adaptation into (i) augmentation-based exploration to identify plausible class candidates, and (ii) fairness-driven calibration that adapts text contexts to equalize sensitivity to common visual evidence. This fairness constraint mitigates partial feature obsession and enables effective calibration of text embeddings without relying on entropy reduction. Through extensive evaluation, we empirically validate our theoretical motivation and show that FCL achieves competitive adaptation performance relative to state-of-the-art TTA methods across diverse domain-shift and fine-grained benchmarks.

  PublisherOA PDF

• Debias Once for All: A Data-Centric Strategy for Fair Machine Learning

  2026-02-16

  articleOpen accessSenior author

  The increasing use of deep neural networks (DNNs) in high-stakes domains such as hiring, healthcare, and finance has heightened concerns about algorithmic fairness. Because training data can encode historical and societal biases, learned models may exhibit disparate outcomes for underrepresented groups. Prior work is largely model-centric, improving fairness via specialized loss functions or architectural modifications, which can introduce additional training overhead and hinder deployment in modular or rapidly evolving pipelines. We instead study a data-centric alternative: constructing a fair training dataset that promotes equitable behavior without changing the model architecture. We propose FairData, which synthesizes a fair dataset by optimizing a gradient-matching objective that aligns the training dynamics of a randomly initialized model on the synthetic data with those on the original data, while explicitly regularizing for group fairness. The resulting dataset is model-agnostic, lightweight, and remains in the original input space, enabling straightforward reuse across downstream models. Experiments on four benchmark datasets show that FairData consistently reduces group disparities across diverse architectures while maintaining competitive predictive performance, suggesting fairness-aware dataset optimization as a practical complement to model-specific fairness techniques.

  PublisherDOI

• Cross-Modal Event Encoder: Bridging Image–Text Knowledge to Event Streams

  2026-03-06

  articleSenior author

  We propose an event-centric encoder that extends the CLIP ecosystem to neuromorphic streams, positioning events as a first-class modality in general-purpose multimodal learning. Our approach transfers CLIP’s image–text knowledge directly to the event domain by introducing a lightweight preprocessing pipeline that collapses raw streams into grayscale frames compatible with CLIP’s vision encoder. This simplification preserves CLIP’s scalability and zero-shot capability, while a carefully designed training scheme mitigates catastrophic forgetting through contrastive, consistency, and distributional alignment losses. The resulting encoder achieves competitive performance on object recognition, few-shot learning, and zero-shot anomaly detection, and generalizes to event streams synthesized from videos without additional training. Beyond standalone benchmarks, we demonstrate seamless integration into cross-modal architectures, enabling event–image retrieval and retrieval across sound and depth, thereby broadening CLIP’s ecosystem. While intentionally simple, our representation serves as a transferable starting point, and future extensions with polarity- or temporal-aware encodings could further exploit event-specific characteristics.

  PublisherDOI

Recent grants

• Neurally-Inspired Integration of Communication and Cognitive Computation in Hyperspace

  NSF · $360k · 2023–2026

• UKRI/BBSRC-NSF/BIO: Interpretable and Noise-Robust Machine Learning for Neurophysiology

  NSF · $797k · 2023–2026

• Hyperdimensional Neural Computation for Real-Time Cognitive Learning

  NSF · $300k · 2021–2025

• CPS: Small: Brain-Inspired Memorization and Attention for Intelligent Sensing

  NSF · $500k · 2023–2027

Frequent coauthors

• Tajana Rosing

  74 shared

• Yeseong Kim

  Daegu Gyeongbuk Institute of Science and Technology

  53 shared

• Tajana Rosing

  47 shared

• Saransh Gupta

  40 shared

• Xunzhao Yin

  Zhejiang University

  35 shared

• Zhuowen Zou

  University of California, Irvine

  32 shared

• Cheng Zhuo

  Zhejiang University

  29 shared

• Justin Morris

  California State University, San Marcos

  26 shared

Awards & honors

• 2024 Young Investigator Award
• DARPA Young Faculty Award (2023)