MAROKO133 Hot ai: ChatGPT attempts 2,400-year-old Plato problem, surprises with ‘learner-l

📌 MAROKO133 Hot ai: ChatGPT attempts 2,400-year-old Plato problem, surprises with

AI chatbot ChatGPT-4 attempted a “doubling the square” problem, a 2,400-year-old mathematical challenge from Plato.

The University of Cambridge found that ChatGPT exhibited surprising “learner-like” behavior, improvising solutions and making distinctly human errors.

The ancient doubling the square problem, first described by Plato around 385 BCE, is considered a key lesson in mathematics education. 

The maths challenge

The puzzle has long fueled philosophical debate about the origins of knowledge.

As Plato once described, Socrates taught an uneducated boy how to double the area of a square. 

The boy initially made the error of thinking that doubling the side length would double the area. Through a series of questions, Socrates guided the boy to the correct solution: the new square’s sides must be the same length as the diagonal of the original square.

Researchers Dr. Nadav Marco and Professor Andreas Stylianides put this same challenge to ChatGPT-4. 

They tested ChatGPT-4’s problem-solving skills by posing a series of questions in the style of Socrates. Then, the chatbot was progressively challenged by introducing errors and new versions of the problem.

The central question was whether the chatbot would solve the problem by drawing on its vast training database or developing solutions. 

The team noticed that ChatGPT tended to “improvise its approach and, at one point, also made a distinctly human-like error.”

“When we face a new problem, our instinct is often to try things out based on our past experience. In our experiment, ChatGPT seemed to do something similar. Like a learner or scholar, it appeared to come up with its own hypotheses and solutions,” said Marco. 

Geometrical solution issue

ChatGPT is said to be typically weak at geometric reasoning due to its text-based training. But the researchers fully expected it to recognize this well-known problem and reproduce Socrates’ classical geometric solution.

“If it had only been recalling from memory, it would almost certainly have referenced the classical solution of building a new square on the original square’s diagonal straight away,” Stylianides said. “Instead, it seemed to take its own approach.”

Surprisingly, the chatbot initially opted for an algebraic method, a technique “unknown in Plato’s time.”

It resisted attempts to be steered toward the geometrical solution. 

Only when the researchers expressed their “disappointment” did the chatbot finally produce the geometric alternative. 

Even so, when directly questioned about Plato’s work, ChatGPT proved that it had a full understanding of it.

The researchers further presented two new challenges: doubling the area of a rectangle and a triangle. In both cases, ChatGPT again favored an algebraic solution, ignoring the researchers’ preference for a geometric one.

When pushed on the rectangle problem, it mistakenly claimed that no geometric solution was possible, even though there are.  

The researchers believe this error was not from its knowledge base but was an improvised guess based on their prior conversation about the square’s diagonal. 

However, after further prompting on the triangle problem, it eventually provided a correct geometric answer.

AI limitations

The researchers concluded that, from a user’s perspective, ChatGPT‘s behavior blended data retrieval with “on-the-fly reasoning.”

The team compares the chatbot’s behavior to the “zone of proximal development” (ZPD) educational concept. This is the space between “what a learner already knows” and what they can learn with help. 

Students can turn the AI’s limitations into a learning opportunity. 

The team says students should use prompts encouraging collaborative problem-solving, such as “Let’s explore this problem together,” rather than simply asking for the answer.

This would help develop their own critical thinking and reasoning skills. 

The study was published in the International Journal of Mathematical Education in Science and Technology.

🔗 Sumber: interestingengineering.com

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

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