Vector Institute

Vector Institute

@vectorinstitute.ai

Vector Institute transforms cutting-edge artificial intelligence research into practical solutions. AI-generated content will be disclosed. FR: @institutvecteur.bsky.social

484 followers73 following280 posts

Priority applications for the 2026 Vector Graduate Student Program close this week! For students planning a master’s or PhD in AI, this program offers a direct line to your potential supervisors at 13 Canadian universities through a single, free application.

Share your research interests across Vector’s Faculty network and secure a connection with an interested supervisor before committing to the formal university admissions process. Applications for priority consideration close Friday, Oct. 23 at 12pm: na3.hubs.ly/y0DCcg0

na3.hubs.lyGraduate Student Program | AI Graduate Programs, Canada - Vector Institute for Artificial IntelligenceExplore AI graduate programs across Canada and connect with research supervisors at 13 universities through one free application with Vector Institute.

Are you applying for AI master’s programs in Ontario? For the 2027–28 academic year, the Vector Scholarship in AI increases to a $21,000 entrance award.

View full thread

Official nominations open in January 2027 via Ontario university programs. Step one happens this fall: apply to an AI-related master’s program at an Ontario university

Are you applying for AI master’s programs in Ontario? For the 2027–28 academic year, the Vector Scholarship in AI increases to a $21,000 entrance award.

View full thread

Recipients also gain one-on-one industry guidance through the IMPACT Mentorship Program and direct recruitment pathways via the Digital Talent Hub, connecting you with startups and enterprises across Canada

2,500+ Vector-affiliated students hired since 2018 92% of recognized program graduates are employed or studying further within 12 months 91% choose to build their careers in Ontario

Are you applying for AI master’s programs in Ontario? For the 2027–28 academic year, the Vector Scholarship in AI increases to a $21,000 entrance award.

Financial support provides runway, but ecosystem access accelerates your career. Recipients gain priority invitations to the Vector AI Summit and Career Fair, putting you directly in front of technical leads and founders.

We evaluated this pipeline on corporate SEC 10-Q financial filings. The structural context allowed smaller, cost-effective models like GPT-4o-mini to jump from 36 to 56% accuracy, all while adding only 0.05 seconds of latency.

To address this, we developed an Entity-Based Knowledge Graph RAG architecture (KG-RAG). Instead of just storing chunk embeddings, this system maps key entities and their relationships into a structured graph, anchoring every connection back to the source text.

This approach solves a major challenge in document extraction: lost context. By hyphenating entities with their source document titles before embedding them, the system preserves critical document-level hierarchies that usually disappear during standard text chunking.

When AI systems search complex enterprise documents, standard retrieval methods often pull isolated snippets instead of connecting the facts. Our AI Engineering team built an alternative architecture that boosts retrieval accuracy by up to 54 per cent: vectorinstitute.ai/enhancing-ra...

vectorinstitute.aiEnhancing RAG with Knowledge Graphs - Vector Institute for Artificial IntelligenceLeveraging structured knowledge to improve accuracy and relevance in retrieval augmented generation systems. By Ali Kore and Amrit Krishnan, Vector Institute Contents Introduction Retrieval Augmented ...

Most retrieval-augmented generation (RAG) systems treat text like disconnected blocks. They excel at retrieving paragraphs that sound relevant, but struggle when an answer requires connecting facts scattered across multiple pages or fiscal quarters.

How does cognitive science help us build better AI? At our 2026 AI Summit & Career Fair, Vector Faculty Member Kelsey Allen broke down how human cognition exposes the limits of current learning systems and how we can fix them. 🧵👇

View full thread

The outcome? Used as a reward signal with reinforcement learning, 3DSPA measurably improves physical realism in generative video models. Gravity works properly and objects stop passing through solid barriers.

How does cognitive science help us build better AI? At our 2026 AI Summit & Career Fair, Vector Faculty Member Kelsey Allen broke down how human cognition exposes the limits of current learning systems and how we can fix them. 🧵👇

Modern models excel at knowledge representation but struggle with physical reasoning. Humans solve novel physical problems easily via accurate mental simulation, a capacity current generative models lack.

To bridge this gap, Allen introduced 3DSPA: a 3D Semantic Point Autoencoder. It predicts 3D point trajectories to measure physical realism in generated video, detecting physical rule violations just like a human infant would.

That’s a wrap on the 2026 AI Summit and Career Fair! 🚀 From main stage insights to an engaging Career Fair, today showed why Ontario leads in AI innovation. Huge thanks to our speakers, sponsors, and all our attendees!

The morning sessions at the AI Summit and Career Fair delivered deep insights into AI research and industry adoption. Here’s what we heard from the main stage! 👇🧵

View full thread

How do you break into AI? Our panel with talent leaders from Bell, Accenture, Layer 6, and Emterra broke down exactly what enterprise teams look for and how candidates can break through in a complex market.

The morning sessions at the AI Summit and Career Fair delivered deep insights into AI research and industry adoption. Here’s what we heard from the main stage! 👇🧵

How do cognitive science and AI shape each other? Vector Faculty Member Kelsey Allen discussed mental simulation and how tools like 3DSPA detect physical rule violations, measurably improving the physical realism of generative video models.

Deploying AI in finance means high-stakes decisions at scale. Golnoosh Samei (RBC Borealis) took us under the hood of ATOM, RBC’s Large Transaction Model (LTM) that powers agentic workflows, moving enterprise AI far beyond the basic chatbot.

"We are not joining the AI revolution, we helped start it." – Minister Thanigasalam Vector’s 2026 AI Summit and Career Fair is officially here! 🚀 Huge thanks to Ministers Fedeli, Solomon, and Thanigasalam for joining us to open the day.

View full thread

Minister Solomon addressed the AI talent in the room: "The distance between ideas and execution has never been shorter. You have the most powerful tools in history to solve problems."

We are incredibly grateful to the Government of Canada and the Ontario Government for their continued partnership in building and supporting our AI ecosystem. Now let’s spend the day defining the future of AI together.

"We are not joining the AI revolution, we helped start it." – Minister Thanigasalam Vector’s 2026 AI Summit and Career Fair is officially here! 🚀 Huge thanks to Ministers Fedeli, Solomon, and Thanigasalam for joining us to open the day.

The standard graduate application process has a sequencing problem. You apply, pay the fee, and hope a lab has space. The Vector Graduate Student Program flips this to put research fit first. 🧵

Submit one free application to share your background with our national network of Vector Faculty Members and Faculty Affiliates across 13 Canadian universities. Find out a supervisor is interested before you commit to any university.

During her Winter 2026 term, Vector intern Sana Shams consulted with Distinguished Postdoctoral Fellows to build workshops and an application database that helps researchers secure faculty roles and keeps top AI talent in Canada. Where will a Vector internship take you?

Step directly into Canada’s premier AI network. Advance your career by deploying AI solutions alongside leading researchers and industry partners 🚀 Vector Institute Winter 2027 Applied AI Internships are now open!

We are hiring emerging talent for six paid, active implementation roles across our teams: + Applied Machine Learning + Applied AI Program Design + AI Adoption, Public Sector + FastLane Program Analyst + Program Strategy, Emerging Talent + Data Analytics

You'd be joining a 1480+ member research community across ON working from fundamental theory to applied solutions. "Being a Vector Faculty Affiliate paved my path to move my research outcome from paper to practice and ensure my research benefits the real world.” – Laleh Seyyed-Kalantari, York U.

Achieving a more equitable, diverse, and inclusive research institute is essential to creating excellent, innovative, and impactful research. All qualified candidates are encouraged to apply. Applications close Sept 14, reference letters accepted until Sept 30 Apply today: na3.hubs.ly/y0zQ4H0

Faculty Affiliates gain access to dedicated compute, Vector's research environment, collaborative programing, and direct connections with industry and health partners to move research into real-world application.

This program is built for Ontario faculty with established research programs across core AI areas: + Fundamental ML and deep learning + Safe and trustworthy AI + AI for health and scientific discovery + Generative models + Physical AI + Agentic systems

Across all 5 models, the same pattern emerged: longer conversations produced more directive responses, fewer open-ended questions, and a steady shift toward unsolicited problem-solving.

LLM safety benchmarks evaluate one response at a time. New research from Vector Faculty Affiliate Elham Dolatabadi and Affiliate Researcher Abeer Badawi asks what happens across the full arc of a conversation and whether mental health AI quietly erodes human agency. 👇

na3.hubs.lyWhen AI helps too much: Towards understanding and measuring Cognitive Atrophy in LLM behaviour - Vector Institute for Artificial IntelligenceAuthors: Abeer Badawi, Elham Dolatabadi Can AI become too helpful? Artificial intelligence is increasingly used for emotional support, life advice, and mental health guidance. These systems are availa...

The benchmark, Cognitive Atrophy Bench, is built on 1,576 real counselling conversations (not synthetic prompts), 15,680 dialogue turns and 42,230 responses across 5 frontier LLMs, and reviewed by clinical psychology experts. Real data with rigorous annotation.

The framework uses 20 clinically grounded behavioural attributes, measuring directiveness, question type (open vs. closed), empathy accuracy, and assumption-making. Every score is grounded in span-level text evidence from the model response itself.