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

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: chain.zhang@jiqizhixin.com

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

📌 TOPINDIATOURS Hot ai: HMND 01 Alpha: UK firm to debut world’s fastest-developed

The HMND 01 Alpha prototype marks a major milestone in humanoid robotics, arriving after just seven months of development.

Built by Humanoid, the wheeled humanoid represents what the company describes as the shortest development cycle yet for a humanoid robot.

Humanoid has already completed its first commercial proofs of concept with global industrial and automotive technology companies.

The company is actively showcasing HMND 01 Alpha at its CES booth from January 6 to January 9, demonstrating the robot in a near-production industrial environment.

HMND 01 Alpha is performing a real-world factory task: autonomously picking metallic bearing rings from cluttered bins.

The demonstration reflects production conditions rather than a controlled lab setup.

HMND 01 Alpha stands 220 centimeters tall and moves on a wheeled base. It reaches speeds of up to 4.47 miles per hour.

The robot carries bimanual payloads of up to 33 pounds and lifts heavier objects when they remain closer to its body.

Its reach spans from floor level to two meters high.

Alpha accesses shelf depths of up to 60 centimeters. This range allows it to pick goods from ground-level bins and elevated storage locations.

The humanoid features 29 active degrees of freedom, excluding end-effectors. AI-driven, end-to-end reasoning powers its motion and task execution.

Operators can equip Alpha with either a 12-DOF five-fingered hand or a 1-DOF parallel gripper, depending on dexterity or load requirements.

A sensor-rich head supports perception. Alpha uses 360-degree RGB cameras and two depth sensors to understand cluttered industrial environments in real time.

Autonomous bin picking

At CES, Alpha is operating live at the Schaeffler Group booth. The robot picks unsorted bearing rings and places them onto a buffer table. In production settings, that buffer feeds directly into a ball-bearing assembly line.

Andreas Zeug of Schaeffler described the workflow in an interview with Interesting Engineering. “It’s press one button, and everything is automated from there.”

He explained the industrial relevance of the task. “The robot takes rings from an unsorted bin and feeds them to a buffer table, which in a real factory leads into a ball-bearing assembly line.”

Zeug emphasized the system’s readiness. “The system is entirely autonomous and completely trained.”

Why wheels, not legs

Schaeffler approaches humanoids as both a user and a supplier. “We are users of humanoids, and we also develop and supply parts for them,” Zeug said.

“We supply actuators, bearings, and transmissions that go into these robots.”

The company evaluates multiple platforms. “We benchmark different humanoid robots to see which fits best for different automation needs.”

Alpha’s wheeled design reflects factory realities.

“For industrial use, we have completely even floors, so we don’t need legs,” Zeug said. “A wheel-based system is more stable and easier to integrate from a safety perspective.”

Looking ahead, Schaeffler expects rapid iteration. “Form factor, reliability, and repairability are the biggest changes going forward,” Zeug added. “Beta will be much more compact and closer to production-ready.”

Scale follows next. “Gamma is when we start producing at scale.”

Humanoid supports that roadmap with $50 million in founder-led capital.

Its 200-person team includes alumni from Apple, Tesla, Google, Boston Dynamics, Sanctuary AI, and Nvidia.

For ongoing news, in-depth reporting, and key developments from CES 2026, read the IE team’s coverage here.

🔗 Sumber: interestingengineering.com