Today, essentially all computing is done with Boolean logic and binary numbers, requiring all manner of error correction codes to keep the most significant bits straight. But in the realm of machine learning, what if instead every bit was equally (in)significant? Consider two randomly generated strings thousands of zeros and ones long, about 50/50 split. Normally, about half the bits will be different between the two strings. If the number of differences is less than half though, the strings as a whole must be related. Even a large number of errant bits will not change the strings’ similarity. This creation and measurement of similarity between very large strings of bits (or numbers) is called hyperdimensional computing (HDC). 

Though not quantum mechanical, a kind of superposition of information across each of the thousands of bits enables machine learning capabilities even Deep Neural Networks (DNN) struggle with, such as learning from a single example or combining the learning from several different sensor modalities. One can even dynamically reprogram swarms of robots in a contested environment – the new program intelligible from a single binary string even with 30-50% noise.

What if we started training artificial neural networks (ANN) to learn and recall these strings: not memorizating or autoencoding but rather “resonating” like a tuning fork? This “resonance” training paradigm suddenly enables even more capabilities not possible with traditional DNNs today such as learning new categories “on the fly” without retraining the whole network from scratch. Such attractor state networks (ASN) lend themselves to modular design, allowing new categories to be added or shared amongst other ASNs. This combining of HDC with “resonance” ANNs is a potential game changer for machine learning, establishing mathematically rigorous engineering design rules instead of the unintelligible black boxes of DNNs today.

Hyperdimensional computing can be implemented in traditional computing hardware, but large size, weight, and power savings are possible by including memristors crossbars into the ML ensemble. The Pipeline for Novel Neuromorphic Hardware (NeuroPipe) Applied Research for Advancement of Priorities (ARAP) is working to bring not only novel neuromorphic computing hardware like memristors to the field but also develop the next generation of machine learning algorithms such as these.


NeuroPipe ARAP Hardware & Algorithms

Monitoring of maritime objects, particularly dark vessels, at a global scale is a challenging task. Synthetic aperture radar (SAR) is a valuable data source because it offers day/night and all-weather situational awareness unlike passive sources such as electro-optical imagery or data sources requiring vessel cooperation such as automatic identification system (AIS) or vessel monitoring system (VMS) data. Deep learning-based computer vision approaches to analyzing SAR imagery have historically been limited by a scarcity of high-quality labeled imagery. To address this problem and to stimulate academic and commercial development of SAR-focused object detectors, the xView3 dataset and challenge was recently introduced to provide annotated imagery for training and evaluation drawn from the ongoing Sentinel-1 satellite SAR mission operated by the European Space Agency. While the xView3 dataset is an important contribution to SAR computer vision research, this is a temporary and limited solution. As novel examples are seen in the wild and as new SAR sensors become available, existing object detection models will need to be retrained with relevant new examples. Given a pool of unlabeled images, the task of selecting new images to label is traditionally assigned to human experts or chosen randomly. As labeling resources are limited, particularly with SAR which requires specialized expertise and ancillary data, it is important to select the most informative images to label so that the model steadily improves as new examples are encountered. Active learning (AL) seeks to automate and optimize this data selection process. Active learners provide an acquisition function which selects a subset of examples for labeling from the set of unlabeled data according to some criteria such as uncertainty, disagreement, diversity, expected error reduction, etc. In this work, we propose to evaluate state-of-the-art AL algorithms for the task of maritime object detection in SAR imagery using the xView3 dataset. To the best of our knowledge, no publicly available work has compared different AL approaches for SAR-based computer vision. We compare their learning curves for different dataset slices (e.g., near-shore, open-ocean, island, etc.) as well as for different object detection models to assess the transferability of AL algorithms to new data and model architectures. Given the high resolution and volume of satellite SAR imagery, we additionally report processing time for each algorithm. We identify gaps in current AL solutions for SAR computer vision and propose solutions to address these shortcomings to improve their utility in practical settings.

Active learning for maritime object detection in synthetic aperture radar imagery

Modern radar simulation development primarily uses raw data collects from blue forces placed in intelligence databases that have not yet been thoroughly vetted and distilled into a usable data product.  These resources often have redundant, ambiguous, or missing information describing modes and operation of the system that must be preprocessed by those local to the organization developing the simulations rather than the Intelligence Communities (ICs) themselves.  The ICs’ process to generate a Tier I or II product takes considerable time and doesn’t always meet the timeline requirement to support the Warfighter in a dynamically changing Electronic Warfare (EW) threat environment.  Hence, it is required for personnel to utilize data-driven tools leveraging Artificial Intelligence (AI) and Machine Learning (ML) to down-select redundant or ambiguous information and impute missing information to construct quick-reaction models at a relatively high frequency.  This effort employs clustering algorithms to find redundant or ambiguous modes for a particular emitter and then constructs the radar simulation based on the consolidation of modes.  The quality is comparable to the traditional manual construction of simulations in many cases and drastically increases the speed and quantity at which simulations are processed.  Missing parameter data is addressed by utilizing a decision-tree based regressor that imputes the missing parameters based on training data taken from other sources that could provide inference (i.e. other radars).  Lowering the variance of these estimates requires execution on data at fairly low levels describing the radar system, but has potential to initially bridge parameter gaps not yet filled. 

Machine learning applications for quick-reaction development of radar simulations with imperfect data

Automated spectrum spectrum management requires real-time feedback from sensors in the sensors in the networknetwork. This real-time feedback requires a large backhaul as the network scales. We employ machine learning classifiers at the edge to reduce the backhaul reduce the bandwidth requirements of the network.

Machine Learning and Radio Frequency Spectrum Management

The explosion of the number of connected devices in the last 10 years has led to a considerable increase in the nature and scope of Artificial Intelligence at the Edge (EdgeAI). Furthermore, hardware, software, and mission use cases have come together to drive several advances in AI supporting Department of Defense applications. During this talk we will describe how Deloitte views EdgeAI as a domain, deployment considerations for EdgeAI technology, and how EdgeAI can impact the challenges faced by the Navy. 

Edge technologies are characterized by a combination of sensor and computation that are networked to achieve an effect. Adding Machine Learning as a software component can give these systems the ability to deliver insights in millisecond time scales. The combination of Edge technologies and deployed Machine Learning models forms an EdgeAI system. 

The Department of Defense has made considerable investment in Artificial Intelligence, and the Department of the Navy has contributed to that effort significantly. Myriad use cases in the core mission domain of the US Navy exist for EdgeAI systems. For example, computer vision models deployed to wearable technologies can provide immediate “walk-through” assistance to Sailors afloat as they work on shipboard weapons systems and power plants. Unmanned systems can similarly be equipped to find and highlight potential issues during maintenance inspections using an entire sensor array. 

Several considerations become important when conducting AI/ML development for Edge applications. Our team uses the latest compute stack for training and deploying models to Edge environments, and we will highlight some of those considerations during this talk. Finally, we will show work that our team has done to address use cases of interest to the Navy mission, demonstrating the art of the possible in the EdgeAI space.

Deployment considerations for EdgeAI Technology

Advancements in computer vision,
machine learning and deep learning have
evolved to a state where real-time
detection and performance exceeds the
human. SeeByte's approach to automatic
target recognition in forward-looking sonar
(FLS) and electro optical camera (EO)
utilizes a convolution neural network
(CNN) and single-shot detector backbone
(SSD).

SeeByte ATR performance for detection and classification subsea objects in sonar and video

Mitigation of range-restricted communication blocks in a distributed swarm is of interest to the Navy, due to the existence of data sharing limitations in all types of sensors and missions carried out in communications-denied environments. If a distributed swarm is affected by a communications block or delay, individual members should still be able to cooperatively identify, evaluate, and act as a team to preserve mission success based on a number of critical parameters guiding the scenario. In our approach, individual members of a heterogeneous swarm in a communication blocked environment construct their own perceived Common Operating Picture (COP). We modified a previously proposed task allocation algorithm to account for incomplete member COPs, and implemented a communications-limiting algorithm in simulation space that emulates a real-world communications block. This block impacts the decision making and asset prioritization of the task allocation algorithm, and our work contributes to a better understanding of task allocation and swarm regrouping in a limited communications environment. Results suggest that compensating for a swarm’s incomplete COP in a limited communications environment achieves a higher likelihood of mission success, and that a quasi-centralized approach for task dissemination ultimately performed more reliably than a fully decentralized approach.

Performative Analysis of Sensor Limitations in Communications Denied Autonomous Swarms

Current approaches for handling machinery control system (MCS) faults on unmanned surface vessels (USVs) are based on guidance and thresholds selected by subject matter experts. While this can be a robust method, it requires significant human input and validation testing in expected operating conditions. To provide vessels with greater autonomy, data driven approaches for MCS faults are of interest. Data driven approaches based on AI/ML, besides saving both time and money, may provide a constructive means to learn from past experience, learn new insights, and optimally handle faults based on the current environment and conditions (e.g. vessel speed, payload, or sea state). One significant challenge with using AI/ML approaches for fault handling is the issue of class balance – in general, one may have limited realizations or samples of a given fault. In order to positively utilize data from previous successful test events, this work proposes the casting the problem of fault detection as an outlier detection problem. By building a healthy state model, detection of deviations from the nominal state provide a means to recognize new or novel operating conditions or the failure or degradation of a given subsystem. By quantifying and tracking the healthy state, long-term trends and typical behavior can be identified. These metrics could also be used to schedule inspections and maintenance based on the condition of the vessel as opposed to predetermined ones. With most outlier detection methods, typically a given threshold probability needs to be selected. Optimal selection of this threshold requires observations of such an outlier. Given that many of the sensors used to trigger hull, mechanical, and electrical (HM&E) faults are sensors with stochastic components (e.g. temperature, pressure, etc.), this work focuses on providing the probability of observing a given sensor reading or fault based on prior measurements. Quantile regression forests (QRF) provide a means to identify outliers, produces measurement statistics, and handle the randomness associated with HM&E sensors. Given time, other AI/ML approaches may be compared to QRF. As many vessels have separate MCS for each subsystem, data engineering is needed to aggregate, synchronize, and decimate data as a preprocessing step. With no further preprocessing, preliminary work has found QRF to be useful to identity component failure. Future efforts plan to scale the training test set to aid in generating long-term trend and statistics and establishing a healthy baseline.

Application of AI/ML to Improve Unmanned Surface Vessel Machinery Control System Fault Handling

Swarm Robotics, which studies collective behaviors of large populations of interacting robots with simple embodied cognition, is an ideal testbed for studying the cultural evolution of language. Our goal is to expand existing simulations in order to explore a dominant theory in evolutionary linguistics—the self-domestication hypothesis—which suggests that increased prosociality led to the sophistication of the cultural niche, enabling the complexification of languages. In our model, robots in multiple nests are engaged in a foraging task (i.e., gathering resources in their environment) while playing a language game. Crucially, we include two novel features: (1) robot individuation: robots have a partner-specific memory, keeping track of the outcomes of past interactions with specific robots; (2) parameterizable prosociality: robots’ tendency to interact is based on experience: successful communication between robots reduces their aggression toward each other and increases their chances of interacting again. First, we show that these manipulations lead to the formation of a classic “in-group bias” where robots favor interaction with some robots over others - a bias which is highly common in social animals in nature but that was so far absent from swarm robotics models. Second, we show that prosociality provides an evolutionary advantage, with higher prosociality values resulting in the collection of more resources. Finally, we show that prosociality modulates the effect of physical distance on lexical convergence, such that low values of innate prosocially lead to more stable sub-swarm divergence, even in relative proximity (i.e., different nests robustly converge on different lexical variants despite being close to each other).

First Steps Toward a Swarm Robotics Model of Self-Domestication and Language Evolution

Reconfigurable spherical robot No.3 can transform from the spherical shape into 3 legs walking using linear actuator. No.3 can walk on flat floor and cross barrier using butterfly walking gait and crawl-kick walking gait.The butterfly walking gait can achieve speed of 5.33 cm/s for flat floor and 1.57 cm/s for crossing the barrier. The new crawl-kick walking gait can achieve speed of 6.15 cm/s for flat floor and 1.51 cm/s crossing the barrier. No.3 can complete function of autonomous transformation, movement on flat floor and cross barrier using 3 legs walking gait. 


A Three-Legged Reconfigurable Spherical Robot No.3

Swarm Robotics, which studies collective behaviors of large populations of robots with simple embodied cognition, is an ideal testbed for studying the cultural evolution of language. Our goal is to expand existing simulations in order to explore a dominant theory in evolutionary linguistics - the self-domestication hypothesis - which suggests that reduced reactive aggression led to the sophistication of the cultural niche, enabling the complexification of languages. 

Current swarm robotics models, however, lack several crucial features that are considered prerequisites for this process. Robots are typically collaborative, homogenous, with little to no memory. To mimic self-domestication, an evolutionary advantage for prosociality needs to be introduced, and robots need to be treated as distinct individuals. 

In our model, robots explore their environment while playing a language game. Crucially, we include two novel features: (1) robot individuation: robots have a partner-specificity memory, keeping track of the outcomes of past interactions with specific robots; (2) parameterizable prosociality: robots’ tendency to interact is based on experience: successful communications between robots reduce their aggression toward each other and increase their chance to interact again. With these manipulations, community structure should emerge (robots favoring interaction with some robots over others, i.e., “in-group” bias), and different lexicons should stabilize in different sub-groups in the swarm despite their physical proximity. 

Future steps will introduce an underlying task requiring robots to cooperate using their shared language to achieve a goal. In this context, more complex language structures should evolve, demonstrating the positive feedback loop between increased self-domestication and language complexity.

A swarm robotics model of the cultural evolution of language

In this paper, a hybrid Otsu and improved Particle Swarm Optimization (PSO) algorithm is presented to deal with multilevel color image thresholding problem, named APSOW. In APSOW, the historical information represented by the local best solutions found so far by PSO population are permuted among the current population, using a randomized greedy process. APSOW also implements a weedout operator to prune the worst individuals from the population. The proposed APSOW is compared to other hybrid EAs and Otsu approaches from literature (include standard PSO model) through twelve benchmark color image problems, showing its potential and robustness.

Underline digital video library

NeuroPipe ARAP Hardware & Algorithms

Active learning for maritime object detection in synthetic aperture radar imagery

Machine learning applications for quick-reaction development of radar simulations with imperfect data

Machine Learning and Radio Frequency Spectrum Management

Deployment considerations for EdgeAI Technology

SeeByte ATR performance for detection and classification subsea objects in sonar and video

Performative Analysis of Sensor Limitations in Communications Denied Autonomous Swarms

Application of AI/ML to Improve Unmanned Surface Vessel Machinery Control System Fault Handling

First Steps Toward a Swarm Robotics Model of Self-Domestication and Language Evolution

A Three-Legged Reconfigurable Spherical Robot No.3

A swarm robotics model of the cultural evolution of language

A Multilevel Thresholding Approach Based on Improved Particle Swarm Optimization for Color Image Segmentation

Stay up to date with the latest Underline news!

PRESENTATIONS

CONFERENCES

COMPANY

RESOURCES