预先布线的静态视觉受体场,用于环境不可知性的感知
Minjun Kang1, Seungdae Baek2, Se-Bum Paik1,2
1Department of Brain and Cognitive Sciences, Korea Advanced Institute of Science and Technology, Daejeon 34141, Republic of Korea.
Patterns (New York, N.Y.)
|February 23, 2026
概括
生物大脑适应不断变化的环境,与传统的深度神经网络 (DNN) 不同. 这项研究表明,DNN中的静态视觉过器能够实现强大的持续学习和环境无关的对象识别.
科学领域:
- 计算神经科学是一种计算神经科学.
- 人工智能的人工智能是人工智能.
- 计算机视觉 计算机视觉 计算机视觉
背景情况:
- 生物大脑对动态环境具有显著的适应能力.
- 传统的深度神经网络 (DNN) 难以应对域移动,限制了它们在现实世界中的应用性.
- 生物大脑中的早期视觉通路开发出静态受体场,有助于环境不可知识别.
研究的目的:
- 调查静态,硬连接的受体场是否可以改善深度神经网络的域概括.
- 为了确定模仿生物视觉处理是否增强人工系统中的持续学习.
- 探索固定早期层过器对物体识别强度的影响.
主要方法:
- 在DNN的早期层中实现了预先开发的Gabor过器,在训练过程中保持它们的静态.
- 在显著的域名转移下训练并测试网络.
- 将修改后的DNN与传统的DNN的概括能力和表示性质进行了比较.
主要成果:
- 使用静态Gabor过器的网络表现出强大的持续学习和跨域的概括.
- 静态过器防止了局部纹理偏差,促进了类似于灵长类动物的基于形状的感知.
- 修改后的DNN实现了通用的隐性空间表示,不同于传统的DNN捕获域特定变异.
结论:
- 静态的,有线的受体场是动态环境中可靠的持续学习的关键生物策略.
- 将生物启发的静态过器纳入DNN可以提高它们对域名转移的稳定性.
- 这种方法为实现更具适应性和通用性的人工智能系统提供了一条途径.
相关概念视频
01:24Vision
60.6K
Vision is the result of light being detected and transduced into neural signals by the retina of the eye. This information is then further analyzed and interpreted by the brain. First, light enters the front of the eye and is focused by the cornea and lens onto the retina—a thin sheet of neural tissue lining the back of the eye. Because of refraction through the convex lens of the eye, images are projected onto the retina upside-down and reversed.
60.6K
01:20Anatomy of the Eyeball
10.1K
The eye is a spherical, hollow structure composed of three tissue layers. The outer layer — the fibrous tunic, comprises the sclera — a white structure — and the cornea, which is transparent. The sclera encompasses some of the ocular surface, most of which is not visible. However, the 'white of the eye' is distinctively visible in humans compared to other species. The cornea, a clear covering at the front of the eye, enables light penetration. The eye's middle...
10.1K
01:32The Retina
77.0K
The retina is a layer of nervous tissue at the back of the eye that transduces light into neural signals. This process, called phototransduction, is carried out by rod and cone photoreceptor cells in the back of the retina.
77.0K
01:26Visual System
2.0K
Light enters the eye through the cornea, a transparent, dome-shaped surface covering the surface of the eyeball that helps to direct and focus incoming light. This light is then channeled toward the pupil, an adjustable opening whose size is controlled by the iris. The iris, a pigmented muscle, regulates the amount of light entering the eye by contracting or dilating the pupil, thereby ensuring optimal light levels for clear vision.
Once through the pupil, the light passes through the lens, a...
Once through the pupil, the light passes through the lens, a...
2.0K
01:22Photoreceptors and Visual Pathways
9.6K
At the molecular level, visual signals trigger transformations in photopigment molecules, resulting in changes in the photoreceptor cell's membrane potential. The photon's energy level is denoted by its wavelength, with each specific wavelength of visible light associated with a distinct color. The spectral range of visible light, classified as electromagnetic radiation, spans from 380 to 720 nm. Electromagnetic radiation wavelengths exceeding 720 nm fall under the infrared category,...
9.6K
01:24Color Vision
1.7K
Color perception begins in the retina, the light-sensitive layer at the back of the eye. Two main theories explain how colors are seen: the trichromatic theory and the opponent-process theory. The trichromatic theory, proposed by Thomas Young in 1802 and extended by Hermann von Helmholtz in 1852, suggests that color vision is based on three types of cone receptors in the retina. These cones are sensitive to different but overlapping ranges of wavelengths corresponding to red, blue, and green.
1.7K