MAROKO133 Update ai: Which Agent Causes Task Failures and When?Researchers from PSU and Du

📌 MAROKO133 Update ai: Which Agent Causes Task Failures and When?Researchers from

Share My Research is Synced’s column that welcomes scholars to share their own research breakthroughs with over 1.5M global AI enthusiasts. Beyond technological advances, Share My Research also calls for interesting stories behind the research and exciting research ideas. Contact us: [email protected]

Meet the authors
Institutions: Penn State University, Duke University, Google DeepMind, University of Washington, Meta, Nanyang Technological University, and Oregon State University. The co-first authors are Shaokun Zhang of Penn State University and Ming Yin of Duke University.

In recent years, LLM Multi-Agent systems have garnered widespread attention for their collaborative approach to solving complex problems. However, it’s a common scenario for these systems to fail at a task despite a flurry of activity. This leaves developers with a critical question: which agent, at what point, was responsible for the failure? Sifting through vast interaction logs to pinpoint the root cause feels like finding a needle in a haystack—a time-consuming and labor-intensive effort.
 
This is a familiar frustration for developers. In increasingly complex Multi-Agent systems, failures are not only common but also incredibly difficult to diagnose due to the autonomous nature of agent collaboration and long information chains. Without a way to quickly identify the source of a failure, system iteration and optimization grind to a halt.
 
To address this challenge, researchers from Penn State University and Duke University, in collaboration with institutions including Google DeepMind, have introduced the novel research problem of “Automated Failure Attribution.” They have constructed the first benchmark dataset for this task, Who&When, and have developed and evaluated several automated attribution methods. This work not only highlights the complexity of the task but also paves a new path toward enhancing the reliability of LLM Multi-Agent systems.

The paper has been accepted as a Spotlight presentation at the top-tier machine learning conference, ICML 2025, and the code and dataset are now fully open-source.

Paper：https://arxiv.org/pdf/2505.00212
Code：https://github.com/mingyin1/Agents_Failure_Attribution
Dataset：https://huggingface.co/datasets/Kevin355/Who_and_When
 
 
Research Background and Challenges
LLM-driven Multi-Agent systems have demonstrated immense potential across many domains. However, these systems are fragile; errors by a single agent, misunderstandings between agents, or mistakes in information transmission can lead to the failure of the entire task.

Currently, when a system fails, developers are often left with manual and inefficient methods for debugging:
Manual Log Archaeology : Developers must manually review lengthy interaction logs to find the source of the problem.
Reliance on Expertise : The debugging process is highly dependent on the developer’s deep understanding of the system and the task at hand.
 
This “needle in a haystack” approach to debugging is not only inefficient but also severely hinders rapid system iteration and the improvement of system reliability. There is an urgent need for an automated, systematic method to pinpoint the cause of failures, effectively bridging the gap between “evaluation results” and “system improvement.”

Core Contributions
This paper makes several groundbreaking contributions to address the challenges above:
1. Defining a New Problem: The paper is the first to formalize “automated failure attribution” as a specific research task. This task is defined by identifying the failure-responsible agent and the decisive error step that led to the task’s failure.
2. Constructing the First Benchmark Dataset: Who&When : This dataset includes a wide range of failure logs collected from 127 LLM Multi-Agent systems, which were either algorithmically generated or hand-crafted by experts to ensure realism and diversity. Each failure log is accompanied by fine-grained human annotations for:
Who: The agent responsible for the failure.
When: The specific interaction step where the decisive error occurred.
Why: A natural language explanation of the cause of the failure.

3. Exploring Initial “Automated Attribution” Methods : Using the Who&When dataset, the paper designs and assesses three distinct methods for automated failure attribution:
– All-at-Once: This method provides the LLM with the user query and the complete failure log, asking it to identify the responsible agent and the decisive error step in a single pass. While cost-effective, it may struggle to pinpoint precise errors in long contexts.
– Step-by-Step: This approach mimics manual debugging by having the LLM review the interaction log sequentially, making a judgment at each step until the error is found. It is more precise at locating the error step but incurs higher costs and risks accumulating errors.
– Binary Search: A compromise between the first two methods, this strategy repeatedly divides the log in half, using the LLM to determine which segment contains the error. It then recursively searches the identified segment, offering a balance of cost and performance.

Experimental Results and Key Findings 
Experiments were conducted in two settings: one where the LLM knows the ground truth answer to the problem the Multi-Agent system is trying to solve (With Ground Truth) and one where it does not (Without Ground Truth). The primary model used was GPT-4o, though other models were also tested. The systematic evaluation of these methods on the Who&When dataset yielded several important insights:
– A Long Way to Go: Current methods are far from perfect. Even the best-performing single method achieved an accuracy of only about 53.5% in identifying the responsible agent and a mere 14.2% in pinpointing the exact error step. Some methods performed even worse than random guessing, underscoring the difficulty of the task.
– No “All-in-One” Solution: Different methods excel at different aspects of the problem. The All-at-Once method is better at identifying “Who,” while the Step-by-Step method is more effective at determining “When.” The Binary Search method provides a middle-ground performance.
 

– Hybrid Approaches Show Promise but at a High Cost: The researchers found that combining different methods, such as using the All-at-Once approach to identify a potential agent and then applying the Step-by-Step method to find the error, can improve overall performance. However, this comes with a significant increase in computational cost.

– State-of-the-Art Models Struggle: Surprisingly, even the most advanced reasoning models, like OpenAI o1 and DeepSeek R1, find this task challenging.- This h…

Konten dipersingkat otomatis.

🔗 Sumber: syncedreview.com

📌 MAROKO133 Hot ai: US firm’s record-breaking drill taps granite 387 feet deep to

A US company has just hosted the first of several live demonstrations showcasing its record-breaking drill that is capable of boring into granite using pure energy.

Massachusetts-based startup Quaise Energy has taken a massive step towards limitless, clean geothermal power with the first live public demonstration of its millimeter wave drilling system.

The demo, held at a granite quarry in Marble Falls, Texas, showed the company’s ability to vaporize rock using high-frequency electromagnetic waves, drilling 387 feet (118 meters) into solid granite without any physical contact.

Dubbed “the first drilling innovation in 100 years,” the technology could unlock access to superhot and superdeep renewable geothermal energy, tapping into the heat stored beneath the Earth’s surface on a scale comparable to fossil fuels.

“Quaise is not a drilling company, it’s an energy company,” Carlos Araque, Quaise Energy’s CEO and co-founder, stated. “We aim to make geothermal the workhorse of the energy transition, and we won’t stop until we succeed.”

First live quarry test

A total of 56 observers witnessed the firm’s system in action at the September 4 event. The demonstration site featured live footage of the drilling hole, real-time data displays, and guided tours of the core components.

This included the gyrotron that generates the millimeter waves, the drilling rig, as well as the operator’s cabin. “At Quaise, we’re rapidly moving from microwaves in the ground to megawatts on the grid,” Matt Houde, Quaise Energy chief of staff and co-founder, elaborated.

According to the company, the attendees even viewed the borehole itself through a camera lowered down the shaft, revealing a smooth transition from surface soil to uniform granite walls.

At approximately 387 feet (118 meters), the hole is the deepest ever drilled with millimeter-waves, which are similar to the microwaves used in an oven. The waves were powerful enough to ablate the pink granite into grey ash at the demo site, where samples were made available during the tour.

This enables a radically different form of drilling, which is potentially capable of reaching the superhot, superdeep geothermal layers miles beneath Earth’s surface.

Breaking drilling records

The Marble Falls demo follows a series of progressively difficult tests. “Last fall we started by drilling four feet into a granite core inside our Houston lab,” Justin Lamb, Marble Falls head of field operations, said.

“Then we moved right outside the lab to drill 10 feet into another core,” he said. In May, the company drilled 40 feet into a granite core using a full-scale oil rig near Houston. The drill was operated by Nabors, one of the world’s leading oil and gas drilling companies.

“We were able to drill that 118 meter hole on our first try,” Lamb explained. “It’s been highly successful, beyond all of our best hopes.”

Although the team wasn’t focused on speed, they still reached drilling rates of up to 16 feet (five meters) per hour through some of the world’s hardest rock. “That’s extremely fast,” Lamb revealed.

Henry Phan, Quaise’s vice president of engineering, noted that typical commercial drilling averages just a tenth of a meter per hour through granite. “Our production goal is eight and a half inches.”

The company’s next goal involves using the millimeter wave technology to drill 10 times deeper and reach a full kilometer within the next months.

“We’ll experiment with various parameters to, for example, control how straight the hole is and see if we can go even faster,” Emilie Williams, Quaise test group manager, said in a press release.

🔗 Sumber: interestingengineering.com