Google, IBM & Meta Certificates – 40% Off
One Coursera Plus subscription covers most Professional Certificates on Coursera.
Unlock All Certificates
Explore the critical question of model reliability in this keynote conference talk that examines how machine learning algorithms, particularly large language models, can deceive users despite appearing highly accurate. Delve into the complexities of assessing model performance in an era where "black box" models are increasingly prevalent in everyday applications. Learn about the various ways models can produce misleading outputs and discover practical approaches for evaluating model trustworthiness. Understand the challenges of validating LLM outputs and gain insights into methodologies for seeing through potential model confusion to uncover reliable results. Examine real-world scenarios where model outputs may not align with expected accuracy and develop strategies for implementing more robust model evaluation practices in your machine learning workflows.
