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

📌 MAROKO133 Breaking ai: Which Agent Causes Task Failures and When?Researchers fro

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 Star Wars-style humanoid robot soldier brings sci-fi

A San Francisco-based robotics company named Foundation has developed what could be the world’s first humanoid robot explicitly designed for warfare. Called the Phantom MK-1, the machine stands 5 feet 9 inches tall, weighs 175 pounds, and can carry loads up to 44 pounds.

The company’s move breaks ranks with most major robotics manufacturers, which have pledged not to weaponize their technology.

A new frontier in military robotics

Across the world, militaries are increasingly integrating artificial intelligence into defense systems. In the Russia–Ukraine war, indigenous drones and unmanned systems are already performing real-time target acquisition and precision strike support. According to one analysis, drones now account for roughly 70-80 % of battlefield casualties, aided by AI-enabled systems.

U.S. and Israel take it one notch further by using AI to track and generate target recommendations, while the People’s Liberation Army (PLA) of China is investing heavily in automated target recognition systems powered by satellite imagery and sensor data.

But Foundation’s approach represents a new dimension. Bringing humanoid robotics into military use. “The future of warfare is real-life video games,” said Sankaet Pathak, the company’s CEO, in an interview with News Nation.

“Air, land, sea, all of them would be autonomous. So in the next 10 years, the most likely scenario is you start seeing a lot of active battlefields where the first body in would be humanoids, and then humans to follow after that if needed.”

Notably, South Korea, India, Australia, the U.K., Saudi Arabia, and the UAE are also developing AI-enabled defense technologies, particularly in computer vision applications that help identify and classify battlefield targets.

Built for the battlefield

According to Pathak, the Phantom MK-1 is intended to act as a ground unit, serving as a first line of defense rather than offense. The robot is designed to perform tasks such as reconnaissance and bomb disposal, reducing the risk to human soldiers in high-risk missions.

The company plans to produce 10,000 units by next year, though Pathak clarified that these robots will not be fully autonomous.

He said that there will be a human operator behind the wheel to take the final judgment to either fire a weapon or not, emphasizing the company’s commitment to keeping “a human in the loop” for mission-critical decisions. While AI will handle tasks like path tracking and trajectory calculations, human oversight will remain essential.

Pathak described the Phantom MK-1 as “faster, stronger, and weaponized to be deadlier,” noting that its design prioritizes durability for deployment in harsh environments. The robotics company aims to build “the most rugged humanoid” capable of withstanding the physical stresses of battlefield operations.

Minimal sensors, maximum control

Unlike many modern humanoids, the Phantom MK-1 relies primarily on cameras rather than LiDAR sensors. In an interview with CNET’s Jesse Orrall, Pathak explained that the team intentionally reduced the number of sensors and cables in the robot to simplify data integration and improve reliability.

Orrall noted that sometimes the data between sensors and cameras conflict, then proceeded to test the robot using a VR headset that allowed direct manual control.

The Phantom MK-1 currently uses a mechanical hand with gloves, which Foundation plans to replace with more advanced hardware in future iterations. The robot’s movements are powered by proprietary cycloid actuators, allowing it to move smoothly, quietly, and with notable strength.

Visually, the Phantom MK-1 has a streamlined, all-black body with broad shoulders and a featureless, mask-like face embedded with cameras, reinforcing its utilitarian, machine-first design.

Its long, forward-curving head gives it a faint resemblance to a Star Wars-style battle droid, though the aesthetic here is rooted in functionality and not fiction. The robot’s camera-first build suggests a focus on vision-based situational awareness, a key element for semi-autonomous operations.

Beyond the battlefield

Though the company’s focus on military applications sets it apart, the company also plans to deploy its robots in factories and logistics. Pathak hinted that the technology could one day extend to Mars exploration, showing ambitions beyond Earthly conflicts.

Founded in 2024, it has quickly attracted attention for focusing on military humanoid development. Out of dozens of humanoid robotics firms worldwide, News Nation reported that it is the only company openly building humanoids for combat.

Whether the Phantom MK-1 becomes a new standard in defense technology or sparks a broader debate over weaponized AI remains to be seen. For now, Foundation and the Phantom MK-1 are a striking evolution in the relationship between tech and warfare.

🔗 Sumber: interestingengineering.com