Publications

Nearly 1,000 citations across 30+ publications at top venues including NeurIPS, ICLR, ICML, AAAI, WACV, Advanced Functional Materials, and more. See also my Google Scholar profile.

AI Trust & Deepfake Detection

Research at ScamAI on detecting synthetic media, AI-generated fraud, and evaluating detection systems in real-world conditions.

How well are open sourced AI-generated image detection models out-of-the-box? A comprehensive benchmark study S Ren, Y Zhou, X Shen, K Zewde, T Duong, G Huang, E Wei, J Xue. arXiv:2602.07814, 2026. Benchmarks open-source AI-generated image detectors on diverse generative models, revealing significant gaps in out-of-distribution generalization.

DOCFORGE-BENCH: A Comprehensive Benchmark for Document Forgery Detection and Analysis Z Zhao, W Xia, P Wei, Y Zhang, Y Zhang, J Mo, T Zhang, Y Dai, Z Chen, … arXiv:2603.01433, 2026. Introduces a large-scale benchmark for document forgery detection spanning multiple manipulation types and document categories.

AIForge-Doc: A Benchmark for Detecting AI-Forged Tampering in Financial and Form Documents J Wu, Y Zhou, M Xu, Z Liang, S Ren, J Xue, M Yang, S Chen, J Huan. arXiv:2602.20569, 2026. [1 citation] Targets AI-forged tampering specifically in financial documents with a new benchmark and detection framework.

GPT4o-Receipt: A Dataset and Human Study for AI-Generated Document Forensics Y Zhang, S Ren, A Raj, E Wei, D Ng, A Shen, J Xu, Y Zhang, E Marotta. arXiv:2603.11442, 2026. Studies human vs. AI ability to detect GPT-4o-generated fake receipts, showing humans struggle while specialized detectors can succeed.

Can a Teenager Fool an AI? Evaluating Low-Cost Cosmetic Attacks on Age Estimation Systems X Shen, T Duong, X An, Z Zhao, Z Hu, H Hu, Z Wang, F Guo, S Ren. arXiv:2602.19539, 2026. [1 citation] Demonstrates that simple cosmetic-based attacks can bypass commercial age estimation systems with alarming success rates.

Out of the box age estimation through facial imagery: A Comprehensive Benchmark S Ren, X Shen, A Raj, A Dai, Y Xu, Z Chen, S Wu, C Gong, Y Zhang. arXiv:2602.07815, 2026. [1 citation] Comprehensive comparison of vision-language models vs. traditional architectures for facial age estimation.

Can Multi-modal (reasoning) LLMs detect document manipulation? Z Liang, K Zewde, RP Singh, D Patil, Z Chen, J Xue, Y Yao, Y Chen, Q Liu, … arXiv:2508.11021, 2025. [3 citations] Evaluates whether multimodal LLMs with reasoning capabilities can detect manipulated documents without specialized training.

Can Multi-modal (reasoning) LLMs work as deepfake detectors? S Ren, Y Yao, K Zewde, Z Liang, NY Cheng, X Zhan, Q Liu, Y Chen, H Xu. arXiv:2503.20084, 2025. [12 citations] First systematic evaluation of multimodal LLMs as zero-shot deepfake detectors, revealing both promise and critical limitations.

Do deepfake detectors work in reality? S Ren, D Patil, K Zewde, TD Ng, H Xu, S Jiang, R Desai, NY Cheng, … Proc. 4th Workshop on Security Implications of Deepfakes, 2025. [7 citations] Real-world stress test of deepfake detectors showing dramatic accuracy drops under compression, social media re-encoding, and adversarial conditions.

Deep Learning for Metamaterial & Photonic Design

Inverse design methods for electromagnetic metamaterials, combining deep learning with physics priors.

Deep Inverse Design of Patchy Particles for Mesoscale Assembly of Superlattices PA Lin, S Ren, et al. 2026. Extends inverse design methods from photonics to colloidal self-assembly, designing patchy particle interactions for target superlattice structures.

Comprehensive Overview of Deep Inverse Models in Metamaterials Design Y Deng, S Ren, C Nadell, JM Malof, WJ Padilla. IRMMW-THz, 2025. Survey of the evolving landscape of deep learning approaches for metamaterial inverse design.

Deep Inverse Design of Metamaterials and Metasurfaces Y Deng, S Ren, JM Malof, WJ Padilla. IEEE AP-S/URSI, 2025. Demonstrates end-to-end deep inverse design pipelines for complex metasurface geometries.

Solving Inverse Problems with Deep Learning WJ Padilla, Y Deng, S Ren, J Malof. ACES Symposium, 2025. Overview of deep learning solutions for inverse electromagnetic problems, covering surrogate models and optimization strategies.

Machine learning for engineering meta-atoms with tailored multipolar resonances W Li, H Barati Sedeh, D Tsvetkov, WJ Padilla, S Ren, J Malof, … Laser & Photonics Reviews 18(7), 2300855, 2024. [26 citations] Uses ML to design meta-atoms with precisely controlled multipolar resonance spectra for advanced photonic applications.

Transfer learning for metamaterial design and simulation R Peng, S Ren, J Malof, WJ Padilla. Nanophotonics 13(13), 2323-2334, 2024. [21 citations] Shows that transfer learning dramatically reduces data requirements for training metamaterial surrogate simulators.

Forward and inverse design of artificial electromagnetic materials JM Malof, S Ren, WJ Padilla. Advances in Electromagnetics Empowered by AI and Deep Learning, 2023. [2 citations] Book chapter covering the full pipeline from forward simulation to inverse design of electromagnetic materials.

Machine learning for mie-tronics W Li, HB Sedeh, WJ Padilla, S Ren, J Malof, NM Litchinitser. arXiv:2305.18589, 2023. [5 citations] Applies ML to Mie resonance engineering, enabling rapid design of dielectric nanostructures with tailored scattering.

Informed deep learning in metamaterials O Khatib, S Ren, J Malof, WJ Padilla. ACES Symposium, 2023. [6 citations] Integrates physical priors into deep learning models to improve accuracy and generalization for metamaterial inverse design.

Metamaterial design with physics informed neural networks O Khatib, S Ren, J Malof, WJ Padilla. SPIE Photonic and Phononic Properties of Engineered Nanostructures XIII, 2023. Demonstrates physics-informed neural networks for metamaterial design that respect Maxwell’s equations.

Deep inverse photonic design: A tutorial Y Deng, S Ren, J Malof, WJ Padilla. Photonics and Nanostructures 52, 101070, 2022. [23 citations] Comprehensive tutorial walking through deep learning methods for inverse photonic device design, with code and benchmarks.

Learning the physics of all-dielectric metamaterials with deep Lorentz neural networks O Khatib, S Ren, J Malof, WJ Padilla. Advanced Optical Materials 10(13), 2200097, 2022. [63 citations] Introduces Lorentz neural networks that embed physical Lorentz oscillator models into the network architecture for physically consistent metamaterial predictions.

Inverse deep learning methods and benchmarks for artificial electromagnetic material design S Ren, A Mahendra, O Khatib, Y Deng, WJ Padilla, JM Malof. Nanoscale 14, 3958-3969, 2022. [61 citations] Comprehensive benchmark of deep inverse methods on electromagnetic material design tasks, establishing standardized evaluation protocols.

Blaschke Product Neural Networks (BPNN): A Physics-Infused Neural Network for Phase Retrieval of Meromorphic Functions J Dong, S Ren, Y Deng, O Khatib, J Malof, M Soltani, W Padilla, V Tarokh. ICLR 2022. Novel physics-infused architecture for phase retrieval leveraging Blaschke product structure of meromorphic functions.

Benchmarking Data-driven Surrogate Simulators for Artificial Electromagnetic Materials Y Deng*, J Dong*, S Ren*, O Khatib, M Soltani, V Tarokh, W Padilla, … NeurIPS 2021. [19 citations] Establishes benchmarks for data-driven forward simulators of electromagnetic materials, comparing architectures and training strategies.

Deep Learning the Electromagnetic Properties of Metamaterials — A Comprehensive Review O Khatib, S Ren, J Malof, WJ Padilla. Advanced Functional Materials, 2101748, 2021. [230 citations] Landmark review of deep learning for metamaterials covering forward modeling, inverse design, and generative approaches across the field.

Neural-adjoint method for the inverse design of all-dielectric metasurfaces Y Deng, S Ren, K Fan, JM Malof, WJ Padilla. Optics Express 29(5), 7526-7534, 2021. [116 citations] Applies the neural-adjoint method to design all-dielectric metasurfaces, achieving high-fidelity inverse design without iterative optimization.

Deep learning and inverse design of artificial electromagnetic materials WJ Padilla, Y Deng, S Ren, J Malof. Metamaterials, Metadevices, and Metasystems, SPIE 11795, 2021. Overview of the deep learning inverse design pipeline for artificial electromagnetic materials.

Benchmarking deep inverse models over time, and the neural-adjoint method S Ren, W Padilla, J Malof. NeurIPS 2020. [86 citations] Introduces the neural-adjoint method with boundary loss for inverse modeling, and proposes time-based benchmarking of inverse solvers.

Machine Learning for Exotic Metasurfaces Y Deng, S Ren, K Fan, J Malof, WJ Padilla. IRMMW-THz, 2020. [1 citation] Early demonstration of ML-driven design for exotic metasurface geometries.

Deep inverse design of hydrophobic patches on DNA origami for mesoscale assembly of superlattices PA Lin, S Ren, JC Piland, LM Collins, S Zauscher, Y Ke, G Arya. NeurIPS 2023 Workshop: AI for Accelerated Materials Design. Applies deep inverse design to DNA origami, optimizing hydrophobic patch placement for self-assembly.

Remote Sensing & Energy

Applying computer vision and deep learning to satellite/drone imagery for energy infrastructure mapping.

Segment anything, from space? S Ren, F Luzi, S Lahrichi, K Kassaw, LM Collins, K Bradbury, JM Malof. WACV 2024. [123 citations] Evaluates Meta’s Segment Anything Model on satellite imagery, establishing baselines and revealing domain gap challenges for earth observation.

Closing the Domain Gap — Blended Synthetic Imagery for Climate Object Detection C Kornfein, F Willard, C Tang, Y Long, S Jain, J Malof, S Ren, K Bradbury. NeurIPS 2022 Workshop: Tackling Climate Change with ML. [4 citations] Uses blended synthetic imagery to bridge domain gaps in climate-relevant object detection from overhead imagery.

Automated extraction of energy systems information from remotely sensed data: A review and analysis S Ren, W Hu, K Bradbury, D Harrison-Atlas, LM Valeri, B Murray, … Applied Energy 326, 119876, 2022. [60 citations] Comprehensive review of ML methods for extracting energy infrastructure information from satellite and aerial imagery.

Self-supervised encoders are better transfer learners in remote sensing applications ZD Calhoun, S Lahrichi, S Ren, JM Malof, K Bradbury. Remote Sensing 14(21), 5500, 2022. [21 citations] Demonstrates that self-supervised pretraining outperforms ImageNet-supervised pretraining for remote sensing downstream tasks.

Meta-simulation for the Automated Design of Synthetic Overhead Imagery H Yu, S Ren, LM Collins, JM Malof. arXiv:2209.08685, 2022. [1 citation] Automates the design of synthetic overhead image generation pipelines using meta-learning.

Utilizing geospatial data for assessing energy security: Mapping small solar home systems using unmanned aerial vehicles and deep learning S Ren, J Malof, TR Fetter, R Beach, J Rineer, K Bradbury. ISPRS Int. J. Geo-Inf. 11(4), 222, 2022. [28 citations] Maps solar home systems in developing regions from drone imagery using deep learning to assess energy access.

Machine Learning Methods

Does Deep Active Learning Work in the Wild? S Ren, S Lahrichi, Y Deng, WJ Padilla, L Collins, J Malof. arXiv:2302.00098, 2023. [3 citations] Investigates whether deep active learning delivers on its promise in realistic scientific computing settings with distribution shift.

Mixture Manifold Networks: A Computationally Efficient Baseline for Inverse Modeling GP Spell, S Ren, LM Collins, JM Malof. AAAI 2023. [2 citations] Proposes a computationally efficient inverse modeling approach using mixture of manifold embeddings.

Towards Robust Deep Active Learning for Scientific Computing S Ren, Y Deng, WJ Padilla, J Malof. arXiv:2201.12632, 2022. [1 citation] Identifies failure modes of deep active learning in scientific computing and proposes robustness improvements.

Drowning Detection based on YOLOv8 improved by GP-GAN Augmentation E Wei, S Ren. 2023. [6 citations] Improves drowning detection accuracy by augmenting training data with GP-GAN generated synthetic images.

Physics

Heterogeneity and Memory Effect in the Sluggish Dynamics of Vacancy Defects in Colloidal Disordered Crystals and Their Implications to High-Entropy Alloys CH Chan, Q Huo, A Kumar, Y Shi, H Hong, Y Du, S Ren, KP Wong, … Advanced Science 9(36), 2205522, 2022. [7 citations] Reveals heterogeneous dynamics and memory effects in colloidal crystals, with implications for understanding high-entropy alloys.

Direct Evidence of Void-Induced Structural Relaxations in Colloidal Glass Formers CT Yip, M Isobe, CH Chan, S Ren, KP Wong, … Physical Review Letters 125(25), 258001, 2020. [46 citations] First direct experimental observation of void-induced structural relaxations in colloidal glass formers.