Registre des initiatives en IA

34 initiative(s) courante(s)

• 101

  Creating space-time continuous surface chlorophyll fields for northwest Atlantic using  neural network methods 

  Neural network methods are applied to fill the space-time gaps in satellite remote sensing data of surface chlorophyll for the Northwest Atlantic. The generated dataset will be validated and analyzed.   

  Pilote Sciences des écosystèmes et des océans IA cognitive et générative Analytique prédictive

• 21

  CSAS Knowledge Base Creation

  This solution leverages the Radia AI Assistant developed by the Department of Innovation, Science and Economic Development (ISED). The purpose of the initiative is to create a knowledge base of CSAS publications from the last 25 years. This knowledge base would allow anyone to leverage the AI Assistant to extract trends, data, and information in such a way as to plot new visualizations by unleashing the information held within the publication repository. This initiative has a Statement of Work with ISED and it is being scoped out for amount of work and cost. All indicators to date show that this is a very workable knowledge base development.

  Idéation IA cognitive et générative Gestion des connaissances

• 5

  ITSD knowledge content improvement

  The DFO IT Service Desk is leveraging generative AI (usually Copilot) to enhance its knowledge base articles and Assyst FAQs, which is knowledge intended for clients. This includes creating new knowledge articles, validating and improving existing content, translating materials, and tailoring messaging to different audiences.

  Pilote Dirigeant principal du numérique IA cognitive et générative IA générative (texte, images, code)

• 18

  Classification AI

  Classification AI (Mark Jarvis) to help with the creation of job descriptions/tasks and analyze job descriptions to help with SoMCs

  Idéation Dirigeant principal du numérique IA cognitive et générative IA générative (texte, images, code)

• 22

  AI Model for Administrative Support

  Pilot with select staff using MS Copilot for briefing material preparation, data analysis, writing, and translation. Exploring broader access for process automation. Action plan for this initiative: 1. Governance: Assign project lead, document user access. Regular check-ins with IT and management. 2. Guardrails: Restrict to trained users, apply GC data/privacy standards, conduct ethical reviews and audits. 3. Work pr

  Pilote Secteur des programmes IA cognitive et générative IA générative (texte, images, code)

• 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 IA générative (texte, images, code)

• 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 Gestion des connaissances

• 63

  Factoid Finder Tool

  The Factoid Finder extracts text from machine-readable PDFs to create searchable content libraries. Small Language Models are used to implement semantic search techniques, allowing users to search the library for relevant passages within the PDFs. Small Language Model and Text Embedder: MS Marco Distilbert Dot-v5 Cross-encoder: MS Marco MiniLM-L6-v2

  Pilote IA cognitive et générative Gestion des connaissances

• 71

  News topic detection tool

  Use large language models and topic clustering to read topics emerging in the news and detect which of those topics is relevant to DFO, as well as what knowledge DFO has available to address those topics. Large Language Models

  Idéation Sciences des écosystèmes et des océans IA cognitive et générative Gestion des connaissances

• 77

  Emulator for OPP port models and for relocatable ocean modelling system

  Use machine learning to develop an emulator (or similar) for the OPP port models, which would be used to produce forecasts via inference. The inference is expected to be both faster and computationally cheaper than existing solutions. The training process would use existing model output from multi-year hindcasts. Success with a fixed location would graduate to developing a relocatable emulator TBD; could be U-Net or CNN, or based on GraphCast

  Idéation Sciences des écosystèmes et des océans IA cognitive et générative Simulation et jumeaux numériques

• 78

  Okanagan sockeye life cycle model

  Using LLM (ChatGPT) to support the development of a statistical lifecycle model for Okanagan sockeye. LLMs are not developing the models but are used to brainstorm modelling approaches, debug code, and confirm that written model descriptions in manuscripts are accurate and clear. AI is not being tasked with developing these models. Large Language Models

  Pilote Sciences des écosystèmes et des océans IA cognitive et générative IA générative (texte, images, code)

• 79

  Biological models to support prioritzing salmon stocks under future climates

  This project fits retrospective, life stage-specific models in a data-driven analytical framework to evaluate functional responses to climate variables across the salmon life cycle and make projections for how stocks will respond to climate change scenarios. This modelling approach is intended to be sufficiently flexible to incorporate variability in data quality. Large language models are being used to brainstorm modelling approaches and debug code. AI is not being tasked with developing these models. Large Language Models

  Idéation Sciences des écosystèmes et des océans IA cognitive et générative Simulation et jumeaux numériques

• 80

  Various Built-In AI Features (Co-pilot, ChatGPT, etc)

  Using built-in AI features and functions in regular workflow operations. This includes LLM's such as ChatGPT, Microsoft Copilot, Gemini, and various others that are integrated directly into the applications already currently being used. Large Language Models

  Pilote Sciences des écosystèmes et des océans IA cognitive et générative IA générative (texte, images, code)

• 84

  Daily Use

  Dashboard Development, Used the standalone M365 version of Copilot to assist with coding.

  Pilote Secteur des programmes IA cognitive et générative IA générative (texte, images, code)

• 85

  Daily Use

  Bing CoPilot, ChatGPT-4 Emails, documents, speaking points, ideas

  Pilote Secteur des programmes IA cognitive et générative IA générative (texte, images, code)

• 229

  DFO PB LLM Service

  Provide an approved DFO internal LLM as a service in a PBMM environment. After our team performed a year-long pilot project providing an unclassified internal GenAI models for 14 teams to develop their custom chatbots, we have collected feedback and identified gaps in the availability of GenAI at DFO. One of the biggest gaps was the need for a secure model to be used with private or classified data. Our cloud platform is certified up to PBMM, so we are working to identify, deploy and approve an internal large language model (LLM), as well as make it available to DFO users through API. IT security team is involved in helping us secure and certify the model for use at DFO.

  Pilote Dirigeant principal du numérique IA cognitive et générative MLOps et gestion des modèles

• 114

  InsightBridge: Executive CSAS Summaries

  Uses AI to convert complex, technical CSAS reports into concise, plain-language summaries tailored for executive decision-makers. Enhances leadership understanding and accelerates informed decision-making by providing easily digestible insights. Ensures executives can quickly grasp critical information without sifting through dense documentation.

  Pilote Sciences des écosystèmes et des océans IA cognitive et générative IA générative (texte, images, code)

• 219

  Fish-Habitat Associations

  Extraction of relevant habitat association info and strengths based on vetted literature input to AI for synthesis. Goal to improve equivalency models used in regulatory decisions. Possible partnership with McMaster University.

  Idéation Sciences des écosystèmes et des océans IA cognitive et générative Gestion des connaissances

• 181

  Pacific Aquacutlure Regulations (PAR) Gap Analysis

  Copilot is being leveraged to develop engagement tools for a regulatory consultation that are intuitive, builds on best practices and existing requirements, and adopts a tone and approach that is adequate for the general public.

  Pilote Secteur des programmes IA cognitive et générative IA générative (texte, images, code)

• 170

  NFIS Elver Retrieval-Augmented Generation Pilot with CDOS

  CDOS RAG pilot accept NFIS submission of illegal elver sales proposal. AI model would receive content and asked to summarize or visualize information based on core concepts to draw inference. Pilot only active for 2 months. Governance: Large project of 40 pilots led by CDOS. Assign project lead; document user access. Regular check-ins with IT and management. Guardrails: Restrict to pilot users; apply GC data/privacy standards; conduct ethical reviews and audits. No interne

  Pilote Secteur des programmes IA cognitive et générative Gestion des connaissances

• 168

  Development of bespoke data-limited tools using AI coding.

  The majority of Canadian fish stocks are data limited thus preventing full analytical stock assessments. Data limited tools are therefore the only approach. We are engaged in quick Ai development of custom data limited assessment approaches for exploited fish stocks

  Pilote Sciences des écosystèmes et des océans IA cognitive et générative IA générative (texte, images, code)

• 167

  Network model developing and testing with synthetic AI participants.

  Qualitative network models (QNM) can incorporate uncertainty through an ensemble of rights holder and stakeholder inputs. Before, engaging these communities, QNMs should be tested and this can be done via the development of synthetic AI participants representing viewpoints to engage with potential social uncertainty before actual real world deployment

  Idéation Sciences des écosystèmes et des océans IA cognitive et générative Simulation et jumeaux numériques

• 144

  Factoid Finder Tool

  This completed project used the Factoid Finder extracts text from machine-readable PDFs to create searchable content libraries. Small Language Models are used to implement semantic search techniques, allowing users to search the library for relevant passages within the PDFs.

  Pilote Sciences des écosystèmes et des océans IA cognitive et générative Gestion des connaissances

• 123

  Fish-Habitat Associations

  Extraction of relevant habitat association info and strengths based on vetted literature input to AI for synthesis. Goal to improve equivalency models used in regulatory decisions. Possible partnership with McMaster University.

  Pilote Sciences des écosystèmes et des océans IA cognitive et générative Gestion des connaissances

• 115

  PublicLens: Science in Plain Language

  Leverages AI to transform highly technical scientific and documents into accessible, plain-language summaries for the Canadian public. Fosters transparency and builds public trust by making scientific and government information easily understandable. Promotes civic engagement by breaking down barriers between experts and everyday Canadians.

  Pilote Sciences des écosystèmes et des océans IA cognitive et générative IA générative (texte, images, code)