通过反事实解释增强 - 修复过度自信的分类器
Sumedha Singla1, Nihal Murali1, Forough Arabshahi2
1University of Pittsburgh.
概括
这项研究通过使用反事实解释来减少过度自信,提高了AI模型的可靠性. 改进后的模型更好地识别了不确定的,未发行和新型样本,这对于安全的AI部署至关重要.
科学领域:
- 人工智能的人工智能
- 机器学习 机器学习
- 计算机视觉 计算机视觉
背景情况:
- 过度自信的人工智能模型在医疗保健和自动驾驶等关键应用中构成风险.
- 模型必须准确地反映在分布和分布之外 (OOD) 模糊样本的不确定性.
- 现有的方法在新或边界病例的细微不确定性量化方面扎.
研究的目的:
- 通过利用反事实解释来解决对AI分类器过度信任的问题.
- 提高模型不确定性估计,以提高关键AI系统的安全性.
- 保持预测性能,同时完善不确定性特征.
主要方法:
- 提议微调预先训练的分类器,使用反事实解释器 (ACE) 的增强.
- 评估了检测远离分布 (远离OOD),接近分布 (接近OOD) 和模两可的样本的方法.
- 专注于改善不确定性测量而不牺牲分类准确性.
主要成果:
- 修订后的模型显示,不确定性量化得到了显著改进.
- 有效地识别了模两可的,远远的OOD和近近的OOD样本,准确度更高.
- 与最先进的不确定性估计技术相比,取得了竞争性表现.
结论:
- 反事实解释为纠正过度自信的人工智能分类器提供了一种可行的方法.
- 基于ACE的微调方法提高了模型可靠性,用于现实世界的部署.
- 这种技术对开发更安全,更值得信赖的AI系统充满希望.
更多相关视频
05:47Evidence-based Knowledge Synthesis and Hypothesis Validation: Navigating Biomedical Knowledge Bases via Explainable AI and Agentic Systems
Published on: June 13, 2025
267
08:05A Prediction Error-driven Retrieval Procedure for Destabilizing and Rewriting Maladaptive Reward Memories in Hazardous Drinkers
Published on: January 5, 2018
9.8K
相关概念视频
Hindsight Biases
3.4K
Hindsight bias leads you to believe that the event you just experienced was predictable, even though it really wasn’t. In other words, you knew all along that things would turn out the way they did. Can you relate this to the phrase "Hindsight is 20/20" now?
3.4K
Fundamental Attribution Error
12.9K
According to some social psychologists, people tend to overemphasize internal factors as explanations—or attributions—for the behavior of other people. They tend to assume that the behavior of another person is a trait of that person, and to underestimate the power of the situation on the behavior of others. They tend to fail to recognize when the behavior of another is due to situational variables, and thus to the person’s state. This erroneous assumption is...
12.9K
Confidence Coefficient
7.7K
The confidence coefficient is also known as the confidence level or degree of confidence. It is the percent expression for the probability, 1-α, that the confidence interval contains the true population parameter assuming that the confidence interval is obtained after sufficient unbiased sampling; for example, if the CL = 90%, then in 90 out of 100 samples the interval estimate will enclose the true population parameter. Here α is the area under the curve, distributed equally under...
7.7K
Confirmation Biases
5.5K
The confirmation bias is the tendency to focus on information that confirms our existing beliefs and ignore information that is inconsistent with our expectations. For example, if you think that your professor is not very nice, you notice all of the instances of rude behavior exhibited by the professor while ignoring the countless pleasant interactions he is involved in on a daily basis. Have you ever fallen prey to the confirmation bias, either as the source or target of such bias?
5.5K
Cause and Effect
10.9K
While variables are sometimes correlated because one does cause the other, it could also be that some other factor, a confounding variable, is actually causing the systematic movement in our variables of interest. For instance, as sales in ice cream increase, so does the overall rate of crime. Is it possible that indulging in your favorite flavor of ice cream could send you on a crime spree? Or, after committing crime do you think you might decide to treat yourself to a cone?
10.9K
Interpretation of Confidence Intervals
5.9K
A confidence interval is a better estimate of the population than a point estimate, as it uses a range of values from a sample instead of a single value.
Confidence intervals have confidence coefficients that are crucial for their interpretation. The most common confidence coefficients are 0.90, 0.95, and 0.99, which can be written as percentages–90%, 95%, and 99%, respectively.
Suppose a person calculates a confidence interval with a confidence coefficient of 0.95. In that case, they can...
Confidence intervals have confidence coefficients that are crucial for their interpretation. The most common confidence coefficients are 0.90, 0.95, and 0.99, which can be written as percentages–90%, 95%, and 99%, respectively.
Suppose a person calculates a confidence interval with a confidence coefficient of 0.95. In that case, they can...
5.9K
