### Teaching Models to Forget: A Deep Dive into Minimally Invasive Machine Unlearning A control-oriented framework for practical, sustainable unlearning. Machine learning systems are now deeply embedded in industries where privacy, compliance, and user trust matter — from financial services and healthcare to... [Read more](/content/teaching-models-to-forget-machine-unlearning/index.html) --- ### From Guesswork to Guarantees: A New Framework for Reliable LLM Summarization The problems with large language models are well documented. You’ve seen the headlines: hallucinations, confident errors, and responses you can’t trust. In high-stakes domains like medicine and law, those failures... [Read more](/content/reliable-llm-document-summarization-with-conformal-prediction/index.html) --- ### How Generative AI Is Transforming the Front Office at TD Securities Discover how Generative AI at TD Securities empowers the Front Office. Explore the TDS AI Virtual Assistant, which uses RAG and Text-to-SQL for instant insights. [Read more](/content/generative-ai-td-securities-front-office/index.html) --- ### Trustworthy AI Must Account for Interactions This post is about our paper entitled "Trustworthy AI Must Account for Interactions" presented at the ICLR 2025 Workshop on Bi-directional Human-AI Alignment. Please refer to the full paper for... [Read more](/content/trustworthy-ai-interactions/index.html) --- ### Static Thread Mapping for High Concurrency XGBoost Inference Servers Introduction. With the proliferation of machine learning (ML) models in all aspects of business, model serving has become a significant area of focus. The rise of high-concurrency (near) real-time ML... [Read more](/content/static-thread-mapping/index.html) --- ### DRESS: Disentangled Representation-based Self-Supervised Meta-Learning for Diverse Tasks In this blog, we dive into some puzzling observations in few-shot learning: despite having access to less information, the simple scheme of pre-training and fine-tuning has been shown to support... [Read more](/content/self-supervised-meta-learning/index.html) --- ### Accelerating AI Outcomes Through Research: How Layer 6 is Building the Future with the Vector Institute As Artificial Intelligence (AI) is gaining widespread use, Layer 6 is using the technology to solve real problems and directly shape how teams work, models are built, and innovation is... [Read more](/content/accelerating-ai-outcomes-through-research-with-the-vector-institute/index.html) --- ### Applying RAG in TD’s North American Customer Operations Centre OpenAI launched GPT 3.5 in November 2022, responding with human-like text and inspiring people to dream up new applications for AI. In the world of traditional ML – regression, classification... [Read more](/content/applying-rag-in-tds-north-american-customer-operations-canadian-centre/index.html) --- ### Uncertainty Isn’t Neutral: How Conformal Prediction Can Amplify Disparities In the age of AI-assisted decision-making, ensuring fairness is more important—and more complicated—than ever. From hiring to healthcare, Machine Learning (ML) models are increasingly used to support or even make... [Read more](/content/uncertainty-isnt-neutral-how-conformal-prediction-can-amplify-disparities/index.html) --- ### Troubleshooting Memory Errors in Python Parallel Processing Leveraging multiple cores is essential for accelerating data – intensive tasks like data analysis, machine learning, and numerical simulations. While GPUs offer immense parallel computing power, this post focuses on... [Read more](/content/troubleshooting-memory-errors-in-python-parallel-processing/index.html)