Humanly Certifying Superhuman Classifiers

• URL: http://arxiv.org/abs/2109.07867v1
• Date: Thu, 16 Sep 2021 11:00:05 GMT
• Title: Humanly Certifying Superhuman Classifiers
• Authors: Qiongkai Xu, Christian Walder, Chenchen Xu
• Abstract summary: Estimating the performance of a machine learning system is a longstanding challenge in artificial intelligence research. We develop a theory for estimating the accuracy compared to the oracle, using only imperfect human annotations for reference. Our analysis provides a simple recipe for detecting and certifying superhuman performance in this setting.
• Score: 8.736864280782592
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: Estimating the performance of a machine learning system is a longstanding challenge in artificial intelligence research. Today, this challenge is especially relevant given the emergence of systems which appear to increasingly outperform human beings. In some cases, this "superhuman" performance is readily demonstrated; for example by defeating legendary human players in traditional two player games. On the other hand, it can be challenging to evaluate classification models that potentially surpass human performance. Indeed, human annotations are often treated as a ground truth, which implicitly assumes the superiority of the human over any models trained on human annotations. In reality, human annotators can make mistakes and be subjective. Evaluating the performance with respect to a genuine oracle may be more objective and reliable, even when querying the oracle is expensive or impossible. In this paper, we first raise the challenge of evaluating the performance of both humans and models with respect to an oracle which is unobserved. We develop a theory for estimating the accuracy compared to the oracle, using only imperfect human annotations for reference. Our analysis provides a simple recipe for detecting and certifying superhuman performance in this setting, which we believe will assist in understanding the stage of current research on classification. We validate the convergence of the bounds and the assumptions of our theory on carefully designed toy experiments with known oracles. Moreover, we demonstrate the utility of our theory by meta-analyzing large-scale natural language processing tasks, for which an oracle does not exist, and show that under our assumptions a number of models from recent years are with high probability superhuman.

Related papers

• Can Language Models Learn to Skip Steps? [59.84848399905409]
  We study the ability to skip steps in reasoning. Unlike humans, who may skip steps to enhance efficiency or to reduce cognitive load, models do not possess such motivations. Our work presents the first exploration into human-like step-skipping ability.
  arXiv  Detail & Related papers  (2024-11-04T07:10:24Z)
• The Generative AI Paradox: "What It Can Create, It May Not Understand" [81.89252713236746]
  Recent wave of generative AI has sparked excitement and concern over potentially superhuman levels of artificial intelligence. At the same time, models still show basic errors in understanding that would not be expected even in non-expert humans. This presents us with an apparent paradox: how do we reconcile seemingly superhuman capabilities with the persistence of errors that few humans would make?
  arXiv  Detail & Related papers  (2023-10-31T18:07:07Z)
• Exploration with Principles for Diverse AI Supervision [88.61687950039662]
  Training large transformers using next-token prediction has given rise to groundbreaking advancements in AI. While this generative AI approach has produced impressive results, it heavily leans on human supervision. This strong reliance on human oversight poses a significant hurdle to the advancement of AI innovation. We propose a novel paradigm termed Exploratory AI (EAI) aimed at autonomously generating high-quality training data.
  arXiv  Detail & Related papers  (2023-10-13T07:03:39Z)
• Improving the Efficiency of Human-in-the-Loop Systems: Adding Artificial to Human Experts [0.7349727826230862]
  We propose a hybrid system that creates artificial experts that learn to classify data instances from unknown classes. Our approach outperforms traditional HITL systems for several benchmarks on image classification.
  arXiv  Detail & Related papers  (2023-07-06T14:06:23Z)
• Designing Closed-Loop Models for Task Allocation [36.04165658325371]
  We exploit weak prior information on human-task similarity to bootstrap model training. We show that the use of such a weak prior can improve task allocation accuracy, even when human decision-makers are fallible and biased.
  arXiv  Detail & Related papers  (2023-05-31T13:57:56Z)
• Are Machine Rationales (Not) Useful to Humans? Measuring and Improving Human Utility of Free-Text Rationales [62.02328001381361]
  We show that human utility of existing rationales is far from satisfactory, and expensive to estimate with human studies. We translate this finding into an automated score, GEN-U, that can help improve LMs' ability to generate rationales with better human utility.
  arXiv  Detail & Related papers  (2023-05-11T19:01:13Z)
• Despite "super-human" performance, current LLMs are unsuited for decisions about ethics and safety [0.0]
  We provide a simple new prompting strategy that leads to yet another supposedly "super-human" result. We find that relying on average performance to judge capabilities can be highly misleading. We also observe signs of inverse scaling with model size on some examples, and show that prompting models to "explain their reasoning" often leads to alarming justifications of unethical actions.
  arXiv  Detail & Related papers  (2022-12-13T00:29:45Z)
• JECC: Commonsense Reasoning Tasks Derived from Interactive Fictions [75.42526766746515]
  We propose a new commonsense reasoning dataset based on human's Interactive Fiction (IF) gameplay walkthroughs. Our dataset focuses on the assessment of functional commonsense knowledge rules rather than factual knowledge. Experiments show that the introduced dataset is challenging to previous machine reading models as well as the new large language models.
  arXiv  Detail & Related papers  (2022-10-18T19:20:53Z)
• Learning to Complement Humans [67.38348247794949]
  A rising vision for AI in the open world centers on the development of systems that can complement humans for perceptual, diagnostic, and reasoning tasks. We demonstrate how an end-to-end learning strategy can be harnessed to optimize the combined performance of human-machine teams.
  arXiv  Detail & Related papers  (2020-05-01T20:00:23Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.