Artificial intelligence (AI)

The rapid adoption and development of artificial intelligence (AI) and Robotic Process Automation (RPA) technologies present significant opportunities to transform how the Government of Canada (GC) delivers services. The integration of AI and RPA solutions offers the potential to automate routine tasks, enhance decision-making, and improve the efficiency of back-office operations.

The AI Strategy for the Federal Public Service outlines a vision for responsible GC adoption of AI for four objectives: deliver world-class services to its clients, protect our people and interests, achieve a more innovative and efficient workplace, and accelerate scientific discovery for the benefit of all..

The GoC is interested in testing AI and RPA technologies that can be adapted to provide solutions for government, capable of automating administrative workflows, performing data capture, improving data processing, and generating recommendations for decision-making purposes.

Definitions:

Artificial intelligence (AI):

A machine-based system that, for explicit or implicit objectives, infers, from the input it receives, how to generate outputs such as predictions, content, recommendations, or decisions that can influence physical or virtual environments. Different AI systems vary in their levels of autonomy and adaptiveness after deployment.

Automated decision-making:

Any technology that either assists or replaces the judgment of human decision-makers. These systems draw from fields like statistics, linguistics and computer science, and use techniques such as rules-based systems, regression, predictive analytics, machine learning, deep learning, and neural nets.

Explainability:

Explainability (also referred to as "interpretability") is the concept that an AI model and its output can be explained to the operator, through a methodological approach consistent with a reasoned principles.

Operational technology (OT):

Hardware and software systems that monitor and control physical devices, processes, and events within an organization.

Robotic process automation (RPA):

A process that can automate a set of routine instructions to perform. RPA allows for simple tasks to be completed at a high volume and speed with few or no errors.

Transparency:

The ability to understand how AI systems operate and make decisions, encompassing concepts such as AI explainability, governance and accountability.

Outcomes:

Innovations must meet at least one of the following outcomes to fulfill the requirements of the SC4 Evaluation Grid:

• [1] RPA solutions that can analyze and automate data processing to perform repetitive administrative, technical, or laboratory tasks;
• [2] AI solutions that can consume data (structured or unstructured) and make decisions to support the following use-cases:
  • [2A] Manage autonomous vehicle traffic and support the navigation of individual or multiple autonomous vehicle(s);
  • [2B] Optimize human resources processes to perform applicant screening and support employee onboarding by generating self-directed training curriculums;
  • [2C] Predict, assess, and model exposure to biothreats to support the management of human, animal or plant disease outbreaks;
  • [2D] Monitor patient health, model outcomes, and support healthcare decision making, or;
  • [2E] Strategic business support solutions to handle queries from a diverse user base of businesses and entrepreneurs that can make recommendations based on feedback from user interactions as well as data insights from disparate datasets to provide robust backend analytics for program and service analysis. The model should leverage previous interactions to optimize future recommendations.
• [3] AI solutions for predictive maintenance or fault detection in OT or IT infrastructure that can:
  • [3A] Automate deployment processes by utilizing AI to support the setup and configuration of mobile IT systems, predicting optimal configurations and automating routine setup tasks;
  • [3B] Enhance mobile infrastructure by using AI-driven real-time data processing and decision-making tools capable of adapting to perform adjustments and optimize performance in response to changing conditions;
  • [3C] Perform remote monitoring, maintenance, and troubleshooting of IT and OT systems to identify potential issues, automate diagnostics, offer recommendations, and perform implementation, or;
  • [3D] Integrate AI to implement security measures and ensure compliance with relevant regulations and standards, including in mobile and deployable contexts.
• [4] Technologies that can perform data-rich capture to convert images and non-text information, including audio and video, into data useable by AI-supported process automation or decision-making tools.

ISC will not accept any of the following (out of scope):

• Solutions that do not have safeguards to secure data, restrict access, and prevent outside intrusion.
• Solutions that do not or are unable to comply with the Directive on Automated Decision Making. See Requirements for more information.

The Government of Canada (GC) is interested in testing enterprise-level information technology system solutions to advance Allied communication capabilities, support the integration of Joint Targeting systems, improve the resilience of the Canadian Armed Forces (CAF) Command and Control infrastructure, and support CAF through the adoption of modern technologies.

Advancements in the next generation of sensing and communication technologies underscore a need to deploy technologies that can support an array of functions to augment intelligence, improve surveillance, and perform reconnaissance.

The GC is interested in testing innovative technologies within four (4) areas of the C4ISR suite of technologies:

• Sensor and surveillance technologies;
• Multi-sensor data fusion technologies;
• Data patterning and exploitation technologies, and;
• Communications technologies.

Definitions:

C4ISR:

Command, Control, Communications, Computers (C4) Intelligence, Surveillance and Reconnaissance (ISR). Advanced C4ISR capabilities provide situational awareness, knowledge of the adversary and environment, and shorten the time between sensing, information synthesis, analysis, and response.

Artificial intelligence (AI):

A subfield of computer science that develops computer programs to behave in a way that would be considered intelligent if observed in a human (e.g., solves problems, learns from experience, understands language, interprets visual scenes).

Dark / Non-cooperative targets:

Targets that do not transmit signals or display markings that allow for easy identification.

Edge / Decentralized computing:

Edge computing moves a portion of storage and compute resources out of the central data center and closer to the source of the data itself. Rather than transmitting raw data to a central data center for processing and analysis, edge computing performs analysis closer to where the data is generated.

Machine learning (ML):

Refers to the technologies and algorithms that enable systems to identify patterns, make decisions, and improve themselves through experience and data. ML is primarily used to process large quantities of data very quickly using algorithms.

Pan-domain:

Solutions capable of operating in multiple domains or environments, including space, air, ground-based, underground or underwater. The term "pan-domain" indicates that solutions are not limited to a single type of environment or application.

Augmented reality (AR):

Technologies that add digital elements over real-world views.

Virtual reality (VR):

A computer-generated environment that simulates a physical presence.

EO/IR sensor system:

Electro-Optical/Infra-Red systems are imaging systems which include both visible and infrared sensors.

Hyperspectral devices:

hyperspectral imaging (HSI) sensors acquire imagery in hundreds of narrow, contiguous spectral bands having a few nanometers spectral width, which facilitates fine discrimination between different features on the Earth's surface

Multi-sensor fusion and analysis:

solutions capable of combining information from several sources to form a unified picture or data set for analysis.

Outcomes:

Innovations must meet at least one of the following outcomes to meet the requirements of the SC4 Call for Proposals Evaluation Grid. Identify in your proposal which outcome(s) you have selected:

• [1] Advanced sensor and surveillance technologies:
  • [1A] Innovative hyperspectral imaging sensor technologies capable of capturing spectral data at different wavelengths of the electromagnetic spectrum;
  • [1B] Innovative visible and near-infrared sensor systems for low-light imaging applications such as target detection and identification at night and under very low illumination conditions;
  • [1C] Innovative solutions to detect and geolocate non-cooperative targets;
  • [1D] Innovative EO/IR sensor systems able to provide high dynamic range imaging for both bright and dim targets across different environments;
  • [1E] Innovative technologies to measures distances and depth that leverage EO/IR sensor systems or neuromorphic cameras, or;
  • [1F] Innovative space surveillance technologies using low size, weight, and power (SWaP) hyperspectral imaging payloads onboard satellites to enable space surveillance and reconnaissance.
• [2] Multi-sensor fusion and analysis technologies with training capabilities:
  • [2A] Innovative solutions that can integrate and visualize data (e.g., optical, radar, infrared), from various sensors into a unified framework to enhance accuracy and enable analysis, consistent with the TAK;
  • [2B] Innovative AI or human-interface adaptive data fusion technologies that can adjust the fusion process based on the changing quality and type of incoming sensor data to perform real-time adaptation, or;
  • [2C] Innovative geospatial analysis tools that integrate and interpret data from various sensors to improve object recognition and object classification to support real-time situational analysis.
• [3] Advanced data exploitation and patterning technologies:
  • [3A] Innovative edge, decentralized, or peripheral computing solutions that process data closer to the source to reduce latency and improve the speed of real-time analysis;
  • [3B] Innovative solutions that can analyze data and establish a hierarchy of importance, based on established mission parameters, to determine the prioritization of information sharing in austere or low-bandwidth environments where complete information transfer may not be possible;
  • [3C] Innovative solutions that can improve upon the exploitation of data acquired by sensor systems to support faster object detection, tracking, and recognition;
  • [3D] Innovative solutions that can exploit hyperspectral image data to generate analytics in near real-time to support pan-domain operational applications;
  • [3E] Innovative and novel modeling techniques to characterize and pattern data acquired by sensor systems, or;
  • [3F] Innovative and novel systems that use AR or VR to overlay multi-sensor data onto physical environments.
• [4] Advanced communication technologies:
  • [4A] Novel high-speed, low-latency communications that facilitate real-time data exchange between control centers and field units in pan-domain environments;
  • [4B] Solutions to establish resilient communication networks used to relay information, improve connectivity and support operational reliability in austere environments with weak or intermittent coverage, or;
  • [4C] Light or laser-based optical communication technologies.

ISC will not accept:

• Solutions proposed under the "[2] Multi-Sensor Fusion and Analysis Technologies" sub-heading that do not have related training components. Training is required to facilitate quick and reliable uptake across the GC; solutions should be intuitive and usable with minimal (<10 hours) training.
• Solutions that do not follow the Tactical Assault Kit (TAK) nomenclature. TAK is a geospatial software application widely used across the GC; adherence to TAK standards is required to ensure useability and interoperability.
• Solutions proposed under "[3] Advanced Data Exploitation and Patterning Technologies" sub-heading that do not map individual sensor inputs when performing data fusion. Individual sensor mapping is required to ensure data is factored appropriately and weighted accurately.
• Solutions proposed under "[4] Advanced Communication Technologies" sub-heading that do not have safeguards to restrict access, secure communications, and prevent intrusion.