夜雨聆风
What Does an AI Do When It Sees an Optical Illusion?
01
PART ONE
Yet in other instances, as recent studies show, these models don't behave like us at all--a fact that reveals key differences between how people and machines "see" the world.
然而，最新研究表明，这些模型在其他场景中的表现与人类截然不同，这一事实揭示了人机“看待”世界方式的关键差异。
Some researchers have found that newly developed vision-language transformer models respond to illusions inconsistently. Sometimes they respond as humans would; in other cases, they provide purely logical and objectively accurate responses. And occasionally they answer with total nonsense, likely the result of hallucination. The motivation behind such studies isn't to prove that humans and AI are alike. One fundamental difference is that our brain is full of nonlinear connections and feedback loops that ferry signals back and forth.
一些研究人员发现，新开发的视觉语言转换模型对视错觉的反应并不稳定。它们有时会像人类一样做出反应；在其他情况下，会给出纯粹理性且客观准确的答案；偶尔还会给出完全荒谬的回答，这很可能是幻觉导致的。这类研究的初衷并非要证明人类和AI是相似的。两者根本差异之一在于，人类的大脑充满了非线性连接与反馈回路，这些回路会来回传递信号。
As our eyes and other sensory systems collect information from the outside world, these iterative networks "help our brains fill in any gaps," says Joel Zylberberg, a computational neuroscientist at York University in Ontario, who wasn't involved in the optical illusion studies.
来自安大略省约克大学、未参与视错觉研究的计算神经科学家乔尔·齐尔伯伯格解释道，当我们的眼睛和其他感觉系统从外界收集信息时，这些迭代网络“帮助大脑填补认知缺口”。
Though some recurrent neural networks have been developed to mimic this aspect of the human brain, many machine-learning models aren't designed to have repetitive, two-directional connections. The most popular generative transformer AI models rely on mathematical functions that are "feed-forward." This means information moves through them in only one direction: from input toward output.
尽管已有循环神经网络被开发出来，试图模仿人脑的这一特性，但多数机器学习模型并未设计具有循环的双向连接。最主流的生成式转换AI模型所依赖的是“前馈”数学函数，这意味着信息只从输入到输出单向流动。
02
PART TWO
Studying how such AI systems react to optical illusions could help computer scientists better understand the abilities and biases of these one-way machine-learning models. It could help AI researchers home in on what factors beyond recurrence are relevant for mimicking human responses.
通过研究AI系统对视错觉的反应，计算机科学家能更好地理解这类单向机器学习模型的能力和偏差。这有助于AI研究员确定除循环网络结构外，还有哪些因素与模仿人类反应相关。
One potential factor is a model's size, according to a team of computer scientists who assessed four open-source vision-language models and presented their findings at a December 2023 conference. The researchers found that larger models, meaning those developed with more weights and variables that determine a response, were more closely aligned with human responses to optical illusions than smaller ones.
一支计算机科学家团队认为，模型规模可能是关键因素之一。该团队评估了四个开源视觉语言模型，并在2023年12月的一次会议上公布了研究成果。研究人员发现，那些权重更多、用于确定反应的变量也更多的大型模型，在应对视错觉方面，比小型模型更贴近人类的反应。
Overall, the AI models the scientists tested were not particularly good at homing in on illusory elements within an image (they were less than 36 percent accurate on average) and only aligned with human responses in about 16 percent of cases on average. Yet the study also found models mimicked humans more closely in response to certain types of illusions than others.
总体而言，受测AI模型对图像中错觉元素的识别能力欠佳（平均准确率低于36%），仅在约16%的情境中与人类反应一致。但研究同时表明，对于特定类型的错觉，模型确实能表现出近似人类的反应。
Asking these models to evaluate perspective illusions, for example, yielded the most humanlike outputs. In perspective illusions, equally sized objects within an image appear to have different sizes when placed against a background that suggests three-dimensional depth.
例如，在要求模型评估透视错觉时，它们给出了最接近人类反应的结果。在透视错觉中，将尺寸相同的物体置于有三维纵深感的背景中，看起来大小不同。
Models were asked to judge the relative size of the silhouette of objects in an image--and the researchers also repeated this test with paired and flipped images to detect any potential right- or left-side bias in the models' responses. If the bot's responses to all questions matched the standard human perception, the study authors considered it "humanlike."
研究人员让模型判断图像中物体轮廓的相对大小，还通过配对镜像图像反复测试，以检测模型是否存在左右侧偏差。若AI对所有问题的回答都与人类标准感知一致，即被认定为“类人反应”。
For one type of prompt, which measured the models' ability to locate objects in an image, the two models tested were up to 75 percent humanlike in responding to perspective illusions. In other tests and for other models, the rates of humanlike responses were considerably lower.
在测试图像中物体定位能力的环节，有两个模型在对透视错觉的反应上与人类吻合度高达75%。然而，在其他测试和其他模型中，类似人类的反应情况的比例则显著下降。
In a separate preprint study released in March, researchers tested the abilities of GPT-4V and Google's Gemini-Pro to evaluate 12 different categories of optical illusions. Those included impossible-object illusions, which are two-dimensional figures of objects that could not exist in three-dimensional space, and hidden-image illusions in which silhouettes of objects are included in an image without being immediately obvious.
在另一项3月发布的预印本研究中，研究人员测试了GPT-4V和谷歌Gemini-Pro应对12种类视错觉的表现。其中包括不可能物体错觉和隐藏图像错觉。比如，不可能物体错觉是指二维图像呈现的物体在三维空间中无法存在；隐藏图像错觉是指图像中包含物体的轮廓，但这些轮廓难以立刻辨认。
In nine out of 12 of the categories, the models were worse at pinpointing what was happening in an illusion compared with people, averaging 59 percent accuracy versus human respondents' 94 percent. But in three categories--colour, angle and size illusions--GPT-4V performed comparably or even slightly better than human reviewers.
在12个类别中，有9类错觉模型的识别准确率低于人类（模型平均59%VS人类94%），但在颜色、角度和大小这三类错觉中，GPT-4V的表现与人类评审相当甚至略胜一筹。
Wasi Ahmad, one of the study authors and an applied scientist in Amazon Web Services' AI lab, thinks the difference comes down to whether analysing the illusions requires quantitative or qualitative reasoning. Humans are adept at both. Machine-learning models, on the other hand, might be less poised to make judgments based on things that can't be easily measured, Ahmad says. All three of the illusion categories which the AI systems were best at interpreting involve quantifiably measurable attributes, not just subjective perception.
该研究作者之一、亚马逊云服务AI实验室的应用科学家瓦西·艾哈迈德认为，关键差异在于分析错觉是否需要定量推理还是定性推理。艾哈迈德说，人类对于两者均擅长，而机器学习模型在判断不易量化的事物时能力较弱。AI在这三类错觉上的最佳表现都涉及可量化特征，而非纯主观感知。
03
PART THREE
To deploy AI systems responsibly, we need to understand their vulnerabilities and blind spots as well as where human tendencies will and won't be replicated, says Joyce Chai, a computer science professor and AI researcher at the University of Michigan and senior author of the preprint presented at the December 2023 conference. "It could be good or bad for a model to align with humans," she says.
密歇根大学计算机科学教授兼AI研究员、该论文的主要作者乔伊斯·柴强调，要负责任地部署AI系统，我们必须充分了解其弱点、盲点以及AI在何种情况下会复制人类认知倾向。她说：“模型与人类认知保持一致可能带来益处，也可能存在弊端。”
In some cases, it's desirable for a model to mitigate human biases. AI medical diagnostic tools that analyse radiology images, for instance, would ideally not be susceptible to visual error. In other applications, though, it might be beneficial for an AI to mimic certain human biases.
在某些应用场景中，我们希望模型能够修正人类认知偏差。例如，在理想情况下，用于分析放射图像的AI医疗诊断工具不应受视觉误差影响。然而在另一些应用场景中，让AI模拟特定的人类认知偏差或许更有益处。
We may want the AI visual systems used in self-driving cars to match human error, Richards points out, so that vehicle mistakes are easier to predict and understand. "One of the biggest dangers with self-driving cars is not that they'll make mistakes. Humans make mistakes driving all the time," he says. But what concerns him about autonomous vehicles is their "weird errors", which established safety systems on the road aren't prepared to handle.
理查兹指出自动驾驶汽车的AI视觉系统若能与人类错误模式相匹配，将更有利于预测和理解车辆错误。“自动驾驶的主要风险并非在于它会犯错，人类开车时本就事故频发。”他说。真正令人担忧的是它们的“异常错误”，现有道路安全系统尚未具备应对这类异常错误的能力。
OpenAI's GPT-4V and other large machine-learning models are often described as black boxes--opaque systems that provide outputs without explanation--but the very human phenomenon of optical illusions could offer a glimpse of what's inside them.
OpenAI的GPT-4V等大型机器学习模型常被称作“黑箱”，但视错觉这种极具人类特征的认知现象，或能为我们提供窥探其内部机制的窗口。
【重点词汇】
motivation n. 动力；激励；积极性
relevant adj. 相关的；适宜的
potential adj./n. 潜在的；潜力
phenomenon n. 现象
【往期回顾】
【END】
文章来源 | Scientific American：What Does an AI Do When It Sees an Optical Illusion?
图片来源 | 搜狗百科
译 者 | 陈凌霄
译 审 | 王春渝、赵敏
复 审 | 李小辉
执行编辑 | 陈凌霄
审核编辑 | 王春渝
