Competency-awareness in trusted computing machine learning



Now apply the same question to artificial intelligence (AI) technology and machine learning? How does a machine know if it's smart enough to do the job? That's exactly what U.S. military researchers are aiming at in the Competency-Aware Machine Learning (CAML) project.






Officials of the Defense Advanced Research Projects Agency (DARPA) in Arlington, Va., issued a broad agency announcement on Tuesday for the CAML project, which focuses on competency-awareness machine learning, where an autonomous system can self-assess its task competency and strategy, and express both in a human-understandable form.

This competency-awareness capability contributes to the goal of transforming autonomous systems from tools into trusted, collaborative partners, DARPA officials say. Competency-aware machine learning will enable machines to control their behaviors to match user expectations and enable human operators quickly and accurately to gain insight into a system’s competence in complex, time-critical, dynamic environments. CAML, in short, seeks to improve human-machine teaming.

State-of-the-art machine learning systems today operate in a complex space, and continuously develop behaviors based on their experiences. Nevertheless, these kinds of smart machines with trusted computingcapabilities are unable to communicate their task strategies, the completeness of their training on a given task, what might influence their actions, or how likely they are to succeed under specific conditions.

Verifying a machine's competence increasingly is unrealistic for human operators. This can be a big problem for the military, where machines often deal with high-stake decisions, and must cope with dynamic, fast-changing conditions.

CAML seeks to improve human-machine teaming capabilities by creating a fundamentally new machine-learning approach, and help human operators choose the right smart machines based on the machines' experience and expertise.

CAML is a four-year program divided into a three-year research first phase, and a one-year technology-demonstration second phase. It focuses on four technology areas: self-knowledge or experience; self-knowledge of task strategies; competency-aware learning; and capability demonstrations.

Self-knowledge of Experiences will develop mechanisms for learning systems to discover conditions encountered during operation, and maintain a memory of experiences.

Self-knowledge of task strategies will enable a machine learning system to analyze its task behaviors, summarize them into generalized patterns, and identify what controls its behavior.

Competency-aware learning integrates component technologies into a competency-aware learning framework that is able to communicate in human-understandable statements. It will conclude with a demonstration on a proposer-provided platform.

Capability demonstrations will show competency-aware machine learning systems on military platforms.