📌 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 Update ai: OpenAI’s Sora 2 Generates Realistic Videos of People Sh

OpenAI has released a new smartphone app — currently invite-only — designed to rival TikTok with an infinite barrage of AI slop.

The app accompanies the company’s latest text-to-video and audio AI generator, Sora 2, which it claims is “more physically accurate, realistic, and more controllable than prior systems.”

A two-minute clip celebrating the announcement was met with predominantly negative reactions, with netizens dismissing it as “unsettling” and “soulless.”

Worse yet, facilitating the AI generation of photorealistic videos could have some concerning implications, especially when it comes to impersonation.

Ironically, OpenAI’s own Sora developer, Gabriel Petersson, demonstrated how easy it was to generate CCTV footage of anyone — in this case, OpenAI CEO Sam Altman — “stealing [graphics cards] at Target.”

The clip shows Altman getting caught by a nearby security guard after trying to walk out of a store with a GPU box — a gag meant to poke fun at the company’s frantic multibillion-dollar bids to secure AI hardware. Specialized AI hardware has become an extremely hot commodity, with AI chipmaker Nvidia announcing a $100 billion partnership with OpenAI just last week.

But the light ribbing of a tech executive aside, the video paints a dystopian picture of a future where anybody could easily be framed for a crime they didn’t commit.

People were quick to point out that Petersson’s gaffe — which was followed by several other videos of Altman sleeping in an office chair, or making people dance on a train platform — felt tone-deaf.

“OpenAI employees are very excited about how well their new AI tool can create fake videos of people doing crimes and have definitely thought through all the implications of this,” Washington Post reporter Drew Harwell posted on Bluesky.

“Every defense attorney now has a pre-written motion when it comes to video evidence, I see,” another user commented.

We’ve already seen instances of law enforcement using AI-powered facial recognition to identify perpetrators, despite glaring inaccuracies in the tech.

As WaPo reported earlier this year, officers in St. Louis used the tech to build a case against an innocent 29-year-old father of four after he was identified by an AI app, despite being warned that it “should not be used as the sole basis for any decision.” While the case was eventually dismissed, experts warn that it could set a worrying precedent.

The use of AI apps to generate transcripts of body cam videos has also raised concerns that the tech could exacerbate existing problems in law enforcement, including racism and sexism.

Now, with the advent of powerful text-to-video AI generators, like Sora 2, it’s becoming even easier to place a target at a crime scene they never visited.

For its part, OpenAI claims that its new app’s “cameo” feature — which allows you to “drop yourself straight into any Sora scene” — will protect regular people from having their appearance show up in AI-generated videos.

“With cameos, you are in control of your likeness end-to-end with Sora,” the company’s announcement reads. “Only you decide who can use your cameo, and you can revoke access or remove any video that includes it at any time.”

“Videos containing cameos of you, including drafts created by other people, are viewable by you at any time,” OpenAI promised.

The company also said that it’s taking “measures to block depictions of public figures” (whether Altman consented to Petersson’s videos remains unclear) and that “every video, profile, and comment can be reported for abuse, with clear recourse when policies are violated.”

It’s too early to tell how all of this will play out. But the sheer fact that the company’s own employees are already demonstrating how easy it is to generate fake videos of innocent people committing crimes doesn’t bode well.

OpenAI has already struggled greatly to implement effective guardrails when it comes to its large language models. It remains to be seen whether Sora will be any different in that respect.

More on OpenAI: OpenAI Ridiculed for Its Latest Cash Grab

The post OpenAI’s Sora 2 Generates Realistic Videos of People Shoplifting appeared first on Futurism.

🔗 Sumber: futurism.com