Regression(回归)
3.Optimization（优化）
step 1：function with unknown
step 2: define loss training data
step 3: optimization
y = 0.1k+0.97x1 achieves the smallest loss L = 0.48k on data of 2017-2020 (training data)
How about data of 2021 (unseen during training)?
feature（特征）
Linear models are too simple... we need more sophisticated modes.
也许显示并不是这样。可能会二次函数。（非线性关系）
[图片上传失败...(image-c9f176-1667375487472)]
Linear models have severe limitation. Model bias
red curve = constant + sum of a set of
[图片上传失败...(image-7cb22c-1667375487472)]
蓝色function 激活函数（sigmoid ReLU）
All Piecewise Linear Curves = constant + sum of a set of 曲折线
Beyond Piecewise Linear?
Approximate continuous curve by a piecewise linear curve.
[图片上传失败...(image-c69ed8-1667375487472)]
How to represent this function?
[图片上传失败...(image-c846f3-1667375487471)]
Sigmoid Function (激活函数)
Hard Sigmoid
[图片上传失败...(image-94f968-1667375487472)]
Different w b c
[图片上传失败...(image-15df29-1667375487472)]
red curve = sum of a set of curve + constant
[图片上传失败...(image-d1449f-1667375487472)]
[图片上传失败...(image-1fe571-1667375487472)]
每一个式子代表不同的function。
New Model: More Features
y = b+wx1
[图片上传失败...(image-7d7880-1667375487472)]
[图片上传失败...(image-b99a92-1667375487472)]
[图片上传失败...(image-b4f1a3-1667375487472)]
[图片上传失败...(image-7fb4a8-1667375487472)]
Function with unknown parameters
[图片上传失败...(image-2ba9d7-1667375487472)]
Back to ML Framework
[图片上传失败...(image-311fad-1667375487472)]
[图片上传失败...(image-d5e7a5-1667375487472)]
Loss:
Loss is a function of parameters L
Loss means how good a set of values is.
[图片上传失败...(image-22596b-1667375487472)]
[图片上传失败...(image-70f81b-1667375487472)]
Optimization of New MOdle
[图片上传失败...(image-ab3b08-1667375487472)]
[图片上传失败...(image-5c9ac5-1667375487472)]
Sigmoid - ReLU
[图片上传失败...(image-cce81f-1667375487472)]
[图片上传失败...(image-373681-1667375487472)]
Activation function
[图片上传失败...(image-293ec4-1667375487472)]
Experimental Results
[图片上传失败...(image-fcbf30-1667375487472)]
Back to ML Framework
[图片上传失败...(image-b1c049-1667375487472)]
[图片上传失败...(image-61e85b-1667375487472)]
Experimental Results
[图片上传失败...(image-5b9e4b-1667375487471)]
3 Layers
Red: real no. of views
Blue: estimated no. of views
[图片上传失败...(image-762e3c-1667375487471)]
[图片上传失败...(image-a76acc-1667375487471)]
Neural Network - This mimics human brains...(???)
[图片上传失败...(image-9422cf-1667375487471)]
Hidder layer
[图片上传失败...(image-442c12-1667375487471)]
[图片上传失败...(image-1db809-1667375487471)]
过拟合
[图片上传失败...(image-1697cf-1667375487471)]
Let's predict no. of views today!
- If we want to select a model for predictiong no. of viwes today, which one will you use?
[图片上传失败...(image-f900f2-1667375487471)]
[图片上传失败...(image-7daef8-1667375487471)]
Park et al. created thin films based on oxygen-deficient tungsten oxide that can modulate the visible and infrared spectral ranges by balance the delicate composition of each constituent.
When different bias voltages are applied to the multilayer DAST films, their refractive index change accordingly to achieve selective transmission of electromagnetic waves.
In this work, we report and numerically investigate the novel multilayer film structure for electrochromic windows based on the electro-optic dielectric material of 4-dimethyl-amino-N-methyl-4-stilbazoliumtosylate (DAST) that they can modulate visible and IR light substantially and independently.
Transparent poly-crystalline spinels protect and defend
P9 2021-Pytorch 教学 part 1
[图片上传失败...(image-f92671-1667375487471)]
