The Pursuit of Precision in AI — Recent advances in foundation models like GPT-3 display remarkable linguistic fluency in generating human-like text. However, despite impressive coherence, these large language models (LLMs) frequently “hallucinate” content that lacks factual accuracy. This propensity for imaginative yet incorrect responses poses particular challenges for deploying LLMs in domains such as healthcare, science…

Explainability is fast emerging as a pivotal concern in artificial intelligence research and deployment. As AI systems grow more omnipresent and autonomous in areas like healthcare, transport, finance, and law, their inner workings seem increasingly impenetrable. However, for such intelligent systems to be trustworthy and ethically integrated into human-centric domains…

Graph neural networks (GNNs) and large language models (LLMs) have emerged as two major branches of artificial intelligence, achieving immense success in learning from graph-structured and natural language data respectively. As graph-structured and natural language data become increasingly interconnected in real-world applications, there is a growing need for artificial intelligence…

When LLMs are used to evaluate qualities like the correctness, accuracy, or relevance of a piece of text, consistency is paramount. If an LLM exhibits inconsistent judgements, then its evaluations become unreliable and untrustworthy. If an LLM evaluates the reasoning quality of arguments, but contradicts itself by rating an invalid…

Personalization broadly refers to tailoring digital experiences to individual users. There are two main approaches to enabling personalization using AI: Explicit personalization involves directly incorporating declared user preferences, profiles, and attributes to steer model outputs. For example, a movie chatbot could generate gendered movie suggestions based on someone specifying they…

Information retrieval (IR) systems have become deeply embedded in modern digital experiences as users increasingly expect quick access to relevant information on demand. From consumer web searches to enterprise knowledge platforms, the ability to synthesize and deliver pertinent data creates tremendous value. A key IR technology gaining traction recently is…

As large technology firms like Google, Meta, and OpenAI continue to make breakthroughs in foundation models for natural language processing, it may seem impossible for other companies, especially smaller private ones, to keep pace. After all, these tech giants are pouring billions into advancing model architectures, recruiting top AI talent…

Artificial intelligence now reshapes markets through revolutionary capabilities like automated language, image recognition, prediction, and optimization. Consumer tech giants like Google, Microsoft, and Tencent pour billions into developing ever-more powerful AI. But the transformational impacts reach into most industries as global companies adopt AI to drive decision-making and efficiency. Economists…

Vast datasets with intricate interconnections exceed limited human capacity for comprehension. With the accelerating pace of change, by the time insights emerge, it may already be too late. Opportunities vanish and small issues snowball as decisions lag behind events. The solution lies not in more consolidated reports, but in…