Registre des initiatives en IA

198 initiative(s) courante(s)

• 26

  Generative AI in intranet Internet content creation

  Integrate Generative AI capacity right into Drupal CMS to help users create draft content, ease translation, generate images, tags and metadata. Technologie already exists, need to be security reviewed, integrated and configured. Challenge, web infrastructure team capacity

  Idéation Dirigeant principal du numérique IA cognitive et générative

• 27

  Turn on AI Auto labeling on all M365 documents and communications channels

  Our E5 M365 licences grant access to AI based tools in Purview to automatically categorized documents and set labels accordingly. Model training would require attention of an Security/IM expert to make perfect. We may want to create specific AI specific labels to avoid conflicts with officials ones.

  Idéation Dirigeant principal du numérique Automatisation et aide à la décision

• 28

  AI-Enabled Electronic Monitoring (EM) Solutions

  The pilot project, in collabration with Fisheries Resources Management, explores the integration of AI-enabled electronic monitoring (EM) solutions to automate the detection and classification of fishing activities using video and sensor data from vessels. This initiative aims to enhance compliance monitoring, reduce manual review time, and support evidence-based fisheries management. Client: Fisheries Resources Management

  Pilote Politiques stratégiques Perception et compréhension

• 29

  AI-Enabled Detection of Watercraft for Aquatic Invasive Species Risk Management

  This pilot project, in collabration with Aquatic Invasive Species Program, develops an AI-powered system to automatically detect and track watercraft at key entry points to prevent the spread of Aquatic Invasive Species (AIS). By enabling early identification of vessels that may carry invasive organisms, the system supports faster inspections and intervention to protect aquatic ecosystems. Initial exploration using sample data is underway, with the project pending ADM approval for installing cameras and data collection. Client: Aquatic Invasive Species Program

  Pilote Politiques stratégiques Perception et compréhension

• 31

  Quality control of oceanographic data - Maritimes Region

  The goal of the project is to assist in the processing and quality control of CTD (oceanographic) data by leveraging machine learning models to predict quality flags for the CTD data products generated by Maritimes region, enabling faster data processing by reducing manual burden. Initial explorations have been conducted to investigate transferability of existing quality control models from Pacific region. Based on investigation results, new models will be trained specific to Maritimes region. Client: Maritimes Region - Ocean Data

  Pilote Politiques stratégiques Automatisation et aide à la décision

• 32

  CTD Anomaly Detection

  The goal of this project is to apply machine learning to detect anomalous oceanographic data, highlighting meaningful physical phenomena in the state of the ocean for closer investigation by scientists. A proof of conept model has been developed to demonstrate feasibility. Further validation is required with the client to verify alignment with the needs of scientists. Client: Pacific Region - Ocean Sciences Division

  Pilote Politiques stratégiques Perception et compréhension

• 33

  Machine-learning-assisted Benthic Marine Life Annotation Tool

  The goal of the project is to improve the efficiency of the image annotation process to reduce manual efforts of the benthic science team and reduce time-to-insight from underwater surveys. Further development is prevented by DFO's IT infrastructure. There are ongoing discussions with CDOS to resolved this but the project is on hold in the meantime. Client: Quebec Region/Pacific Region - Oceans and Ecosystem Sciences Division

  Idéation En pause Politiques stratégiques Perception et compréhension

• 34

  AI-assisted Marine Mammal Annotation Tool

  The goal of the project is to improve the efficiency of the aerial image annotation process to reduce manual efforts of the marine mammal science team and reduce time-to-insight for marine mammal survey flights. This project is currently in progress. Business requirements have been finalized and development has commenced Client: Ecosystems and Oceans Science - Ecosystems Science Directorate

  Pilote Politiques stratégiques Perception et compréhension

• 35

  AI-assisted Solution to Detect Ghost Gear from Side Scan Sonar Images

  The goal of the project is use AI to automate data analysis in detecting Ghost Gear. Client: Ghost Gear Program

  Idéation En pause Politiques stratégiques Perception et compréhension

• 36

  AI-assisted Solution to summarize C&P aerial surveillance videos

  A pilot project to explore the use of AI/ML for automating the identification of relevant segments in EM video and imagery, such as human activity or vessel registration numbers (VRNs). This initiative aims to reduce manual review time and enhance the efficiency of information management in support of fisheries oversight.

  Pilote Politiques stratégiques Perception et compréhension

• 37

  OCDS AI Portal

  The Office of Chief Data Steward (OCDS) currently has in its possession multiple proof-of-concept AI models developed that demonstrates the potential of AI in fisheries and oceans management. Due to a lack of deployment processes, the models currently sits idle in our repository. This initiative aims to develop a portal where the models can be deployed in a proof-of-concept fashion to enhancement DFO's understanding of AI and drive the cultural change needed to promote the adoption of AI at DFO. The portal has been completed and is accessible at: https://ocds-ai-portal.canadacentral.cloudapp.azure.com/ The portal was shared in the DFO kiosk at the 2025 GC data conference for engagement. It has so far welcomed 360 users, received over 2K views an

  Pilote Politiques stratégiques Infrastructure et habilitation

• 41

  Cyber Security Operation

  Cyber Security Operation Centre has deployed AI/ML in MS Defender (Endpoint Detection and Response (EDR) system), MS Sentinel (Security Incident and Event Management (SIEM) system), Data Lost Prevention (DLP), and Exchange Online Protection (EOP) for real-time threat detection.

  Production Dirigeant principal du numérique Automatisation et aide à la décision

• 43

  PSPC Concierge

  In partnership with CDOS and Pay Programs, we are agreeing to test the PSPC chatbot with 50 users, slated to start in June. This test will determine if a chatbot is a viable solution to answer generic pay quests.

  Idéation Ressources humaines et services de gestion Interaction et engagement

• 45

  MyGCHR – Phoenix analysis tool

  Pro B data would be needed, but feed MyGCHR and Phoenix records, but flag those that are incorrect in the hopes of training the AI to recognize future incorrect entries and to help clean up pay files for employees. P&C is in negotiation currently with PSPC to ‘own’ our backlog of Phoenix cases, so this could be a big help.

  Idéation Ressources humaines et services de gestion Automatisation et aide à la décision

• 49

  Salmon Fence Counting​

  A model using computer vision to analyze water stream fence video-footage, to simultanously count the number of and predict the species of salmon.​ YOLO 11 Chinook trained on 735 annotated frames from 63 videos Coho trained on 432 annotated frames from 44 videos Sockeye trained on 1149 annotated frames from 36 videos Chum untrained (annotation of 1368 frames from 43 videos in progress) Pink Humpback

  Pilote Secteur des programmes Perception et compréhension

• 50

  Salmon Data Digitization​ (See B22, B26)

  Digitizing non-machine-readable (handwritten) documents and extracts characters into a digital format so that it is accessible, editable, and useable for analysis. Document Intelligence - Custom Classification and Extraction Model Azure OpenAI GPT4o mini Data and Information from ~45,000 pages have been digitized

  Pilote Secteur des programmes Perception et compréhension

• 51

  EN/FR Translation Model​

  A large language model that is trained on the Translations Bureau’s and DFO's language-characteristics (i.e. grammar, annotation, syntax, words, scientific term) context. The model’s enables translation to be like those by the Translation Bureau. ​​Training on large database of scientific documents (published by Canadian Science Advisory Secretariat) to understand scientific terminology is currently in progress. MADLAD 400 10B (Initially trained on sample of ~100 translated documents, in progress of additional training with published CSAS documents going back to year 2000)

  Pilote Secteur des programmes Interaction et engagement

• 52

  "Chumputer Vision" (see B32) Salmon Scale Aging​

  A model using computer vision to look at images of salmon scale and predict the age of salmon and also monitor its ring growth. CNN and YOLO 9. Currently trained on 1478 images.

  Pilote Secteur des programmes Perception et compréhension

• 53

  LLM Document Chatbot​

  Large language model that can help users ask questions to document sets and get answers, summarize key information, and find data with citations and sourcing., including, from documents with multiple templates using prompt engineering approaches Azure OpenAI GPT4o/4o mini

  Développement Secteur des programmes Interaction et engagement

• 54

  PII & Sensitivity Scores​

  Determines the sensitivity of the content of a document based on a trained model and allows for the redaction of the sensitive material in the document. Presidio

  Idéation En pause Secteur des programmes Automatisation et aide à la décision

• 55

  Natural Language Process (NLP) Chatbot using retrival-augmented-generation (RAG) on DFO PAC Site

  A chatbot that allows users to search DFO PAC website using natural language and return reponses not only on data it was trained on but also incorporating a retrieval of new information found on DFO website as it is updated Azure OpenAI GPT4o/4o mini Text-Embedding-3-Large Open Web UI Crawl4AI (AI-based webscraper)

  Développement Secteur des programmes Interaction et engagement

• 56

  Method Recommendation Engine​

  A large language model reviews a scientist’s objective for collecting data in the field and based on the criteria, the model will provide the best procedure to use for their work

  Idéation Secteur des programmes Automatisation et aide à la décision

• 57

  CSAS Knowledge Graph

  Using LLM to extract ontologies and relationships within published CSAS reports into a Graph RAG so that users can visualize and query relationships between researches

  Idéation Secteur des programmes IA cognitive et générative

• 58

  Phytoplankton image classification model

  Create a neural network model to classify phytoplankton images collected by in-situ phytoplankton imaging sensors. There are two discrete steps to this project: 1) create a library of annotated images for model training; 2) creation and refinement of the CNN model. Convolutional Neural Network, InceptionV3 and ResNet

  Pilote Sciences des écosystèmes et des océans Perception et compréhension

• 59

  Clustering to characterize extreme marine conditions for the benthic region of the Northeastern Pacific continental margin

  We introduce a method for characterizing extremes that uses machine learning to divide the data into regions with relatively consistent environmental conditions (temperature, oxygen, acidity), and define the extremes based on the historical statistics of variability of each of these fields. unsupervised clustering with k-means

  Pilote Sciences des écosystèmes et des océans Automatisation et aide à la décision