AI agents that leverage Large Language Models (LLMs) are increasingly becoming core building blocks of modern software systems. A wide range of frameworks is now available to support the specification of such applications. These frameworks enable the definition of agent setups using natural language prompting, which specifies the roles, goals, and tools assigned to the various agents involved. Within such setups, agent behavior is non-deterministic for any given input, highlighting the critical need for robust debugging and observability tools. In this work, we explore the use of process and causal discovery applied to agent execution trajectories as a means of enhancing developer observability. This approach aids in monitoring and understanding the emergent variability in agent behavior. Additionally, we complement this with LLM-based static analysis techniques to distinguish between intended and unintended behavioral variability. We argue that such instrumentation is essential for giving developers greater control over evolving specifications and for identifying aspects of functionality that may require more precise and explicit definitions.

Autonomous business processes (ABPs), i.e., self-executing workflows leveraging AI/ML, have the potential to improve operational efficiency, reduce errors, lower costs, improve response times, and free human workers for more strategic and creative work. However, ABPs may raise specific concerns including decreased stakeholder trust, difficulties in debugging, hindered accountability, risk of bias, and issues with regulatory compliance. We argue for eXplainable ABPs (XABPs) to address these concerns by enabling systems to articulate their rationale. The paper outlines a systematic approach to XABPs, characterizing their forms, structuring explainability, and identifying key BPM research challenges towards XABPs.

Analytic continuation aims to reconstruct real-time spectral functions from imaginary-time Green's functions; however, this process is notoriously ill-posed and challenging to solve. We propose a novel neural network architecture, named the Feature Learning Network (FL-net), to enhance the prediction accuracy of spectral functions, achieving an improvement of at least $20\%$ over traditional methods, such as the Maximum Entropy Method (MEM), and previous neural network approaches. Furthermore, we develop an analytical method to evaluate the robustness of the proposed network. Using this method, we demonstrate that increasing the hidden dimensionality of FL-net, while leading to lower loss, results in decreased robustness. Overall, our model provides valuable insights into effectively addressing the complex challenges associated with analytic continuation.

It was nothing more than what Brad Stevens termed "a curveball," as it turned out. After an initial false positive COVID test, Evan Fournier turned in a string of negative tests, leading to his first-time availability for the Celtics Monday night against New Orleans. "He will play significant minutes, as he will all the rest of the year," Stevens said of how he planned to begin with the talented wing player, acquired from Orlando at the trade deadline for the since-waived Jeff Teague and two second-round draft picks. "We had an obvious need for another wing that can do what he does, and we're fortunate he's with us, and he's on our team," said the Celtics coach. "So I got a chance to go over to the gym (Sunday) while he was shooting around when we got back and then this morning we went through some stuff prior to our shootaround, we shot around as a team for 30 minutes, so he's gotten the crash course in a very short amount of time. He's been there, done that. He's played against us, you know, tons of times, probably knows our plays as well as anybody, and certainly we just want him to play to his strengths and not worry about anything else."

New Zealand scientists have invented a new volcano alert system that they say could have provided warning ahead of last year's White Island disaster. Twenty-one people died when the country's most active volcano, also called Whakaari, suddenly erupted last December with tourists on it. The new system uses machine learning algorithms to analyse real-time data to predict future eruptions. The research was publish in the journal Nature last week. One of the scientists involved in the project, Shane Cronin from the University of Auckland, told the BBC the current system had been "too slow to provide warnings for people [on] the island." "The current [alert system] collects data in real-time but what tends to happen is that this information gets assessed by a panel and they have an expert process... this all takes a while," he said.

We've read the stories about how artificial intelligence (AI) and machine learning are transforming the way companies approach marketing. But what if the true game-changer in consumer insights will be driven by our own brainwaves? Although it may sound like science fiction, the technology has been around for several years, and some companies are finding ways to use brain data to drive product development and market research. In fact, neuromarketing--which uses brain research to reveal a consumer's subconscious decision-making processes--has been in use for more than a decade. In 2009, PepsiCo's Cheetos used EEGs from the brain to measure consumer response to a "prank" type ad, and learned its focus group wasn't quite forthcoming with its written responses.