United States

NAML 2022

Lt. Gen. Matthew G. Glavy, Deputy Commandant for Information, USMC

keynote

##### For full functionality of Underline, please SHUT OFF your VPN, or use a personal device.


Naval Information Warfare Center Pacific is partnering with AFCEA San Diego chapter to host the workshop.

The NAML workshop is an opportunity for researchers from Navy research and development centers and our partner organizations to exchange technical ideas and to come together as a team.

The goals of NAML are:

To spread awareness of current machine learning research relevant to Naval applications

To connect machine learning researchers with experts in military requirements

To build and strengthen collaborations within the defense research community


<figure class="video_container">
  <iframe src="https://s3.amazonaws.com/pf-user-files-01/u-59356/uploads/2022-03-22/hk13mxo/NAML_ReceptionWithtitle.mp4" width=826  height=526  frameborder="0" allowfullscreen="true"> </iframe>
</figure>

Naval Information Warfare Center Pacific is partnering with AFCEA San Diego chapter to host the workshop. The NAML workshop is an opportunity for researchers from Navy research and development centers and our partner organizations to exchange technical ideas and to come together as a team.

The goals of NAML are:

To spread awareness of current machine learning research relevant to Naval applications

To connect machine learning researchers with experts in military requirements

To build and strengthen collaborations within the defense research community



Artificial Intelligence (AI) Safety is currently an area within the field of computer science with a strong emphasis on developing safe algorithms for autonomous vehicles such as self-driving cars and unmanned aircraft. As AI is integrated into more and more systems, the hazard analysis process has to be expanded to account for potential mishaps involving these new technologies. A better understanding of the capabilities of AI is needed to create safe systems. This work will give an overview and discuss some of the current development in AI Safety across the Department of Defense.

Designing Safety into AI Enabled Systems

One of the major issues surrounding the adoption of machine learning in applications where accuracy is critical is the lack of trust in such systems due to being persistently perceived as black boxes. In this presentation, we discuss the importance of independently measuring the competency of a machine learning system and how that builds user trust. We define competency as the expected performance of a ML system for a given set of environmental conditions. We describe how to measure competency based on environmental conditions, what those conditions are and how they can be defined. We give examples of user-friendly competency statements and define competency metrics. Given that maritime environmental conditions can change rapidly, we also discuss ways to decide how often to update the user on changes in competency. We then introduce BAE System's MindfuL™ software which produces competency statements. We show results for our defined competency metrics, namely that for a given set of environmental conditions considered in a competency statement, the actual performance of the system will fall within the competency statement’s confidence interval over 99% of the time. 

The second part of the presentation will focus on the reliability of a machine learning system in the context of competency and expands on how this helps build user trust. In other words, we consider the system’s ability to behave as expected as based on the stated competency out in the real world. If there is an expected performance and set of resulting actions taken by the system for a given set of conditions, reliability is the likelihood of achieving that expected performance and resulting actions. Finally we conclude with thoughts on additional ways to build user trust via MindfuL™ software.


Measuring the competency and reliability of ML Systems

Traditional software safety techniques rely on verifying software against a deductively defined specification of how the software should behave. In the case of AI systems, however, specifications are often implicit (e.g., in the case of expert systems, linear and higher-order programs, and rules engines, where behavior is induced by describing facts, deduction rules, or objective functions and allowing an inference engine or solver to recombine them according to pre-set combination schemas) or inductively defined (e.g., in machine learning systems, the specification is derived from the data set and perhaps some parameters of the model). Data-driven methods are subject to systematic bias and are epistemically limited: practical datasets cannot describe all possible events in any open-world problem. Traditional software verification and validation approaches may not apply directly to these novel systems, complicating the operation of systems safety analysis (e.g., MIL-STD 882E). However, AI offers advanced capabilities, and these capabilities must be safe to employ. When AI is deployed in a weapon, robot, or planning system, unwanted events may occur. Several techniques can be used to support the evaluation process for understanding and making risk decisions for military employment. This research considers those techniques – including software analysis, simulation environments, and mathematical determinations – and evaluates which are most employable, useful, and correct.

Understanding and Assessing Safety in Artificial Intelligence Systems

A viable approach to algorithm development and acquisition takes into account the efficacy of the development and deployment cycle, the reproducibility of models and predictions, the ease of diagnosing algorithm failures and inadequacies, verifiability of compliance with requirements, scalability of both the evaluation system and the algorithm under evaluation, the ability to collaborate in order to accelerate development, efficiency and accuracy of performance benchmarking, and the difficulty of moving an algorithm from an evaluation to an operational environment. Taken together, these concepts support solid engineering practice and due diligence in algorithm acquisition. When applied to machine learning approaches, they are commonly termed the MLOps (Machine Learning Operations) cycle. They are an integral part of viable acquisition and management of software algorithms. As part of the software life cycle an acquisition process must manage a set of operationally meaningful performance tests and their relationship to versioned software implementations of algorithms. This paper describes the Seascape system and its place in such a process.

At the program level, the acquisition of algorithms takes place at a distance from the development work. Vendors of algorithms may be defense companies, national laboratories, collaborative research partners, or other independent groups. The distance between the vendor of the algorithm and the program manager produces a need for an objective way to assess and compare any algorithms with which they are presented. It is not guaranteed that the individual in charge of acquisition will be an expert in all types of algorithms under consideration, further highlighting the need for an objective framework that permits standardized comparisons. Ideally such a system is highly scalable, allowing for parallel evaluation of multiple algorithms in an efficient fashion. Data should be carefully curated and stored, to provide an objective frame of reference for an algorithm, and the analysis should be done in an automated way, ensuring that all algorithms are assessed in the same way every time they are changed.

Seascape: An Algorithm Benchmarking Framework

In this talk, we tackle a generic n-dimensional numerical computing problem and explore performance and tradeoffs between popular frameworks based on open source jupyter notebook examples. Most data science practitioners perform their work in Python because of its high level abstraction and rich set of numerical computing libraries. But the choice of library and methodology is driven by complexity-impacting constraints like problem size, latency, memory, SWAP, hardware, and others. To that end, we demonstrate that a wide selection of GPU-accelerated libraries (RAPIDS, CuPy, Numba, Dask, cuNumeric), including the development of hand tuned CUDA kernels, are accessible to data scientists without ever leaving Python.

Evaluating Your Options for Accelerated Numerical Computing in Pure Python

Preprocessing EEG data for downstream classification tasks

Cooperative autonomous systems, such as swarms, multi-camera systems, or fleets of self-driving vehicles, can illuminate a given scene by providing multiple modalities and varying viewpoints. This redundancy and superposition of data adds robustness to a system typically burdened with obstructions and unrecognizable, distant objects. Collaborative perception is a key component of cooperative autonomous systems where modalities can include camera sensors, LiDAR, RADAR, and depth images. Meanwhile, the amount of useful information that can be shared between agents in a cooperative system is constrained by current communication technologies (e.g. bandwidth limitations). Recent developments in learned compression can enable the learning of end-to-end cooperative systems using deep learning with compressed communication in the pipeline. We explore the use of a deep learning object detector in a cooperative setting with a learned compression model facilitating communication between agents. To test our algorithm, this research will focus on object detection in the image domain as a proxy for one of the modalities used by collaborative systems.

Bandwidth constrained cooperative object detection in images

Addressing quality-of-service (QoS) requirements for software applications
in undersea acoustic networks is challenging. This is in
part due to the strong dependency of the characteristics on the
acoustic propagation environment and the lack of effective congestion
control protocols for undersea acoustic networks. In this
work we propose a deep reinforcement learning (DRL) based algorithm
for allocating bandwidth resources across different traffic
types. Bandwidth allocations are periodically updated based on
traffic prioritization, QoS delay requirements, and queue occupancy
levels. A DRL agent learns the optimal bandwidth allocation policy
via a Soft Actor-Critic (SAC) algorithm. In order to test the
effectiveness of our proposed algorithm, we developed a weighted
fair queuing (WFQ) module in NS3 to enable testing in a high fidelity
simulation environment. The WFQ module enables the use of
queueing disciplines together with an NS3 library for underwater
acoustic communications and networking. The interaction between
the DRL agent and the simulated environment is enabled by the
NS3-gym framework. Numerical tests illustrate the performance of
the DRL-based policy and a use case for the new WFQ NS3 module.

Active Queue Management in Underwater Network via Deep Reinforcement Learning

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).

Lt. Gen. Matthew G. Glavy, Deputy Commandant for Information, USMC

Downloads

Next from NAML 2022

Designing Safety into AI Enabled Systems

Stay up to date with the latest Underline news!

PRESENTATIONS

CONFERENCES

COMPANY

RESOURCES