Sean McPherson

Dr. Sean McPherson

orcid

Research Focus & Metrics

Academic Career Length

8Years (2019-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
11 March 2026 in Nature Biomedical Engineering
Verified
Journal Article
Closed access
JCJonathan Scott Calvert
Jonathan Scott Calvert
SPSamuel R. Parker
Samuel R. Parker
LGLakshmi N. Govindarajan
Lakshmi N. Govindarajan
RDRadu Darie
Radu Darie
ESElias Shaaya
Elias Shaaya
DBDavid A. Borton
David A. Borton

Verified
Journal Article
Open access
PDF file
LWLauren Wong
Lauren Wong
BMBraeden Muller
Braeden Muller
SMSean McPherson
Sean McPherson
Alan MichaelsAlan Michaels
Alan Michaels
Transfer learning (TL) techniques, which leverage prior knowledge gained from data with different distributions to achieve higher performance and reduced training time, are often used in computer vision (CV) and natural language processing (NLP), but have yet to be fully utilized in the field of radio frequency machine learning (RFML). This work systematically evaluates how the training domain and task, characterized by the transmitter (Tx)/receiver (Rx) hardware and channel environment, impact radio frequency (RF) TL performance for example automatic modulation classification (AMC) and specific emitter identification (SEI) use-cases. Through exhaustive experimentation using carefully curated synthetic and captured datasets with varying signal types, channel types, signal to noise ratios (SNRs), carrier/center frequencys (CFs), frequency offsets (FOs), and Tx and Rx devices, actionable and generalized conclusions are drawn regarding how best to use RF TL techniques for domain adaptation and sequential learning. Consistent with trends identified in other modalities, our results show that RF TL performance is highly dependent on the similarity between the source and target domains/tasks, but also on the relative difficulty of the source and target domains/tasks. Results also discuss the impacts of channel environment and hardware variations on RF TL performance and compare RF TL performance using head re-training and model fine-tuning methods.

Verified
Journal Article
Open access
PDF file
LWLauren Wong
Lauren Wong
BMBraeden Muller
Braeden Muller
SMSean McPherson
Sean McPherson
Alan MichaelsAlan Michaels
Alan Michaels
The use of transfer learning (TL) techniques has become common practice in fields such as computer vision (CV) and natural language processing (NLP). Leveraging prior knowledge gained from data with different distributions, TL offers higher performance and reduced training time, but has yet to be fully utilized in applications of machine learning (ML) and deep learning (DL) techniques and applications related to wireless communications, a field loosely termed radio frequency machine learning (RFML). This work examines whether existing transferability metrics, used in other modalities, might be useful in the context of RFML. Results show that the two existing metrics tested, Log Expected Empirical Prediction (LEEP) and Logarithm of Maximum Evidence (LogME), correlate well with post-transfer accuracy and can therefore be used to select source models for radio frequency (RF) domain adaptation and to predict post-transfer accuracy.

13 May 2024 in arXiv
Verified
Preprint
Open access
PDF file
HVHector A. Valdez
Hector A. Valdez
SMSean McPherson
Sean McPherson
We use positional-unigram byte models along with maximum likelihood for generalized TLS fingerprinting and empirically show that it is robust to cipher stunting. Our approach creates a set of positional-unigram byte models from client hello messages. Each positional-unigram byte model is a statistical model of TLS client hello traffic created by a client application or process. To fingerprint a TLS connection, we use its client hello, and compute the likelihood as a function of a statistical model. The statistical model that maximizes the likelihood function is the predicted client application for the given client hello. Our data driven approach does not use side-channel information and can be updated on-the-fly. We experimentally validate our method on an internal dataset and show that it is robust to cipher stunting by tracking an unbiased $f_{1}$ score as we synthetically increase randomization.

15 August 2019 in arXiv
Verified
Preprint
Open access
PDF file
NWNicholas Weir
Nicholas Weir
DLDavid Lindenbaum
David Lindenbaum
ABAlexei Bastidas
Alexei Bastidas
AEAdam Van Etten
Adam Van Etten
SMSean McPherson
Sean McPherson
HTHanlin Tang
Hanlin Tang
Detection and segmentation of objects in overheard imagery is a challenging task. The variable density, random orientation, small size, and instance-to-instance heterogeneity of objects in overhead imagery calls for approaches distinct from existing models designed for natural scene datasets. Though new overhead imagery datasets are being developed, they almost universally comprise a single view taken from directly overhead ("at nadir"), failing to address a critical variable: look angle. By contrast, views vary in real-world overhead imagery, particularly in dynamic scenarios such as natural disasters where first looks are often over 40 degrees off-nadir. This represents an important challenge to computer vision methods, as changing view angle adds distortions, alters resolution, and changes lighting. At present, the impact of these perturbations for algorithmic detection and segmentation of objects is untested. To address this problem, we present an open source Multi-View Overhead Imagery dataset, termed SpaceNet MVOI, with 27 unique looks from a broad range of viewing angles (-32.5 degrees to 54.0 degrees). Each of these images cover the same 665 square km geographic extent and are annotated with 126,747 building footprint labels, enabling direct assessment of the impact of viewpoint perturbation on model performance. We benchmark multiple leading segmentation and object detection models on: (1) building detection, (2) generalization to unseen viewing angles and resolutions, and (3) sensitivity of building footprint extraction to changes in resolution. We find that state of the art segmentation and object detection models struggle to identify buildings in off-nadir imagery and generalize poorly to unseen views, presenting an important benchmark to explore the broadly relevant challenge of detecting small, heterogeneous target objects in visually dynamic contexts.