Sitemap
A list of all the posts and pages found on the site. For you robots out there, there is an XML version available for digesting as well.
Pages
publications
A Data Science Pipeline for Algorithmic Trading: A Comparative Study of Applications for Finance and Cryptoeconomics
Published in 2022 IEEE International Conference on Blockchain (Blockchain), 2022
Recent advances in Artificial Intelligence (AI) have made algorithmic trading play a central role in finance. However, current research and applications are disconnected information islands. We propose a generally applicable pipeline for designing, programming, and evaluating the algorithmic trading of stock and crypto assets. Moreover, we demonstrate how our data science pipeline works with respect to four conventional algorithms; the moving average crossover, volume-weighted average price, sentiment analysis, and statistical arbitrage algorithms. Our study offers a systematic way to program, evaluate, and compare different trading strategies. Furthermore, we implement our algorithms through object-oriented programming in Python3, which serves as open-source software for future academic research and applications.
Recommended citation: Zhang, Luyao, Tianyu Wu, Saad Lahrichi, Carlos-Gustavo Salas-Flores, and Jiayi Li. "A data science pipeline for algorithmic trading: A comparative study of applications for finance and cryptoeconomics." In 2022 IEEE International Conference on Blockchain (Blockchain), pp. 298-303. IEEE, 2022.
Download Paper
Self-Supervised Encoders Are Better Transfer Learners in Remote Sensing Applications
Published in Remote Sensing, 2022
Transfer learning has been shown to be an effective method for achieving high-performance models when applying deep learning to remote sensing data. Recent research has demonstrated that representations learned through self-supervision transfer better than representations learned on supervised classification tasks. However, little research has focused explicitly on applying self-supervised encoders to remote sensing tasks. Using three diverse remote sensing datasets, we compared the performance of encoders pre-trained through both supervision and self-supervision on ImageNet, then fine-tuned on a final remote sensing task. Furthermore, we explored whether performance benefited from further pre-training on remote sensing data. Our experiments used SwAV due to its comparably lower computational requirements, as this method would prove most easily replicable by practitioners. We show that an encoder pre-trained on ImageNet using self-supervision transfers better than one pre-trained using supervision on three diverse remote sensing applications. Moreover, self-supervision on the target data alone as a pre-training step seldom boosts performance beyond this transferred encoder. We attribute this inefficacy to the lower diversity and size of remote sensing datasets, compared to ImageNet. In conclusion, we recommend that researchers use self-supervised representations for transfer learning on remote sensing data and that future research should focus on ways to increase performance further using self-supervision.
Recommended citation: Calhoun, Zachary D., Saad Lahrichi, Simiao Ren, Jordan M. Malof, and Kyle Bradbury. "Self-supervised encoders are better transfer learners in remote sensing applications." Remote Sensing 14, no. 21 (2022): 5500.
Download Paper
Segment Anything, From Space?
Published in Proceedings of the IEEE/CVF Winter Conference on Applications of Computer Vision (WACV), 2024
Recently, the first foundation model developed specifically for image segmentation tasks was developed, termed the “Segment Anything Model” (SAM). SAM can segment objects in input imagery based on cheap input prompts, such as one (or more) points, a bounding box, or a mask. The authors examined the zero-shot image segmentation accuracy of SAM on a large number of vision benchmark tasks and found that SAM usually achieved recognition accuracy similar to, or sometimes exceeding, vision models that had been trained on the target tasks. The impressive generalization of SAM for segmentation has major implications for vision researchers working on natural imagery. In this work, we examine whether SAM’s performance extends to overhead imagery problems and help guide the community’s response to its development. We examine SAM’s performance on a set of diverse and widely studied benchmark tasks. We find that SAM does often generalize well to overhead imagery, although it fails in some cases due to the unique characteristics of overhead imagery and its common target objects. We report on these unique systematic failure cases for remote sensing imagery that may comprise useful future research for the community.
Recommended citation: Ren, Simiao, Francesco Luzi, Saad Lahrichi, Kaleb Kassaw, Leslie M. Collins, Kyle Bradbury, and Jordan M. Malof. "Segment anything, from space?." In Proceedings of the IEEE/CVF Winter Conference on Applications of Computer Vision, pp. 8355-8365. 2024.
Download Paper
Does Deep Active Learning Work in the Wild?
Published in arXiv, 2024
Deep active learning (DAL) methods have shown significant improvements in sample efficiency compared to simple random sampling. While these studies are valuable, they nearly always assume that optimal DAL hyperparameter (HP) settings are known in advance, or optimize the HPs through repeating DAL several times with different HP settings. Here, we argue that in real-world settings, or in the wild, there is significant uncertainty regarding good HPs, and their optimization contradicts the premise of using DAL (i.e., we require labeling efficiency). In this study, we evaluate the performance of eleven modern DAL methods on eight benchmark problems as we vary a key HP shared by all methods; the pool ratio. Despite adjusting only one HP, our results indicate that eight of the eleven DAL methods sometimes underperform relative to simple random sampling and some frequently perform worse. Only three methods always outperform random sampling (albeit narrowly), and we find that these methods all utilize diversity to select samples - a relatively simple criterion. Our findings reveal the limitations of existing DAL methods when deployed in the wild, and present this as an important new open problem in the field.
Recommended citation: Ren, Simiao, Saad Lahrichi, Yang Deng, Willie J. Padilla, Leslie Collins, and Jordan Malof. "Does Deep Active Learning Work in the Wild?." arXiv preprint arXiv:2302.00098 (2024).
Download Paper
Is Self-Supervised Pre-training on Satellite Imagery Better than ImageNet? A Systematic Study with Sentinel-2
Published in arXiv, 2025
Self-supervised learning (SSL) has demonstrated significant potential in pre-training robust models with limited labeled data, making it particularly valuable for remote sensing (RS) tasks. A common assumption is that pre-training on domain-aligned data provides maximal benefits on downstream tasks, particularly when compared to ImageNet-pretraining (INP). In this work, we investigate this assumption by collecting GeoNet, a large and diverse dataset of global optical Sentinel-2 imagery, and pre-training SwAV and MAE on both GeoNet and ImageNet. Evaluating these models on six downstream tasks in the few-shot setting reveals that SSL pre-training on RS data offers modest performance improvements over INP, and that it remains competitive in multiple scenarios. This indicates that the presumed benefits of SSL pre-training on RS data may be overstated, and the additional costs of data curation and pre-training could be unjustified.
Recommended citation: Lahrichi, Saad, Zion Sheng, Shufan Xia, Kyle Bradbury, and Jordan Malof. "Is self-supervised pre-training on satellite imagery better than imagenet? a systematic study with sentinel-2." arXiv preprint arXiv:2502.10669 (2025).
Download Paper
Advancing Time Series Wildfire Spread Prediction: Modeling Improvements and the WSTS+ Benchmark
Published in arXiv, 2025
Recent research has demonstrated the potential of deep neural networks (DNNs) to accurately predict wildfire spread on a given day based upon high-dimensional explanatory data from a single preceding day, or from a time series of T preceding days. Here, we introduce a variety of modeling improvements that achieve state-of-the-art (SOTA) accuracy for both single-day and multi-day input scenarios, as evaluated on a large public benchmark for next-day wildfire spread, termed the WildfireSpreadTS (WSTS) benchmark. Consistent with prior work, we found that models using time-series input obtained the best overall accuracy. Furthermore, we create a new benchmark, WSTS+, by incorporating four additional years of historical wildfire data into the WSTS benchmark. Our benchmark doubles the number of unique years of historical data, expands its geographic scope, and, to our knowledge, represents the largest public benchmark for time-series-based wildfire spread prediction.
Recommended citation: Lahrichi, Saad, Jake Bova, Jesse Johnson, and Jordan Malof. "Advancing Time Series Wildfire Spread Prediction: Modeling Improvements and the WSTS+ Benchmark." arXiv preprint arXiv:2502.12003 (2025).
Download Paper
Key Considerations for Robust Near Field Response Prediction and Optical Metasurface Inverse Design
Published in SPIE, 2025
In this work, we investigate the problem of real-time design of electromagnetic (EM) metamaterials to achieve custom scattering properties; a type of inverse modeling problem. To address this problem, we investigate a class of DNN-based models that are specially designed to address inverse problems, termed deep inverse models (DIMs). DIMs have recently shown tremendous promise for solving material design problems, however, relatively less work has been done for high-dimensional problems, such as near-field design. In this work, we performed 1500 simulations of a metasurface with a 3x3 array of meta-atom pillars, where we independently and randomly-varied the radii of each pillar and recorded the resulting electric near-field values. We then used this dataset to train and evaluate several data-driven inverse models, including several variations of a recently-successful DIM, termed the Tandem. Our results indicate that the Tandem is capable of making relatively accurate design predictions in this challenging high-dimensional settings, and doing so in real-time (e.g., roughly 4ms). We find that the choice of model architecture significantly impacts the accuracy of the inverse model, and even higher accuracy can be achieved with further improvements to the Tandem’s design.
Recommended citation: Mick, Ethan J., Marshall B. Lindsay, Scott D. Kovaleski, Derek T. Anderson, Saad Lahrichi, Jordan Malof, Steven R. Price, and Stanton R. Price. "Key considerations for robust near-field response prediction and optical metasurface inverse design." In Advanced Optics for Imaging Applications: UV through LWIR X, vol. 13466, pp. 61-73. SPIE, 2025.
Download Paper
Deep Inverse Modeling the Near Field Response of Optical Metasurfaces
Published in SPIE, 2025
In this work, we investigate the problem of real-time design of electromagnetic (EM) metamaterials to achieve custom scattering properties; a type of inverse modeling problem. To address this problem, we investigate a class of DNN-based models that are specially designed to address inverse problems, termed deep inverse models (DIMs). DIMs have recently shown tremendous promise for solving material design problems, however, relatively less work has been done for high-dimensional problems, such as near-field design. In this work, we performed 1500 simulations of a metasurface with a 3x3 array of meta-atom pillars, where we independently and randomly-varied the radii of each pillar and recorded the resulting electric near-field values. We then used this dataset to train and evaluate several data-driven inverse models, including several variations of a recently-successful DIM, termed the Tandem. Our results indicate that the Tandem is capable of making relatively accurate design predictions in this challenging high-dimensional settings, and doing so in real-time (e.g., roughly 4ms). We find that the choice of model architecture significantly impacts the accuracy of the inverse model, and even higher accuracy can be achieved with further improvements to the Tandem’s design.
Recommended citation: Lahrichi, Saad, Ethan J. Mick, Marshall B. Lindsay, Scott D. Kovaleski, Derek T. Anderson, Jordan M. Malof, Stanton R. Price, and Steven R. Price. "Deep inverse modeling the near field response of optical metasurfaces." In Advanced Optics for Imaging Applications: UV through LWIR X, vol. 13466, pp. 40-49. SPIE, 2025.
Download Paper
talks
GeoNet, A Global Dataset and Foundation Model for Deep Learning on Optical Satellite Imagery
Published:
teaching
Advanced Algorithms
Graduate course, University of Montana, Department of Computer Science, 2023
Teaching Assistant for the Advanced Algorithms Class. Delivered 4 lectures on Divide & Conquer techniques.
Machine Learning
Graduate Course, University of Montana, Department of Computer Science, 2023
Teaching Assistant for the Machine Learning Class.
Deep Learning
Graduate Course, University of Montana, Department of Computer Science, 2024
Teaching Assistant for the Deep Learning Class. Designed 3 lab sessions, walking students through the complete deep learning pipeline (preprocessing, fine-tuning, evaluation, and visualization)