-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy pathfermi.html
336 lines (250 loc) · 20.1 KB
/
fermi.html
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
<!DOCTYPE html>
<html lang="en">
<head>
<meta charset="UTF-8">
<meta name="viewport" content="width=device-width, initial-scale=1.0">
<title>Fermi Questions</title>
<link rel="stylesheet" href="styles.css">
<!-- Latest compiled and minified CSS -->
<link rel="stylesheet" href="https://cdn.jsdelivr.net/npm/[email protected]/dist/css/bootstrap.min.css"
integrity="sha384-BVYiiSIFeK1dGmJRAkycuHAHRg32OmUcww7on3RYdg4Va+PmSTsz/K68vbdEjh4u" crossorigin="anonymous">
<!-- Latest compiled and minified JavaScript -->
<script src="https://cdn.jsdelivr.net/npm/[email protected]/dist/js/bootstrap.min.js"
integrity="sha384-Tc5IQib027qvyjSMfHjOMaLkfuWVxZxUPnCJA7l2mCWNIpG9mGCD8wGNIcPD7Txa"
crossorigin="anonymous"></script>
</head>
<body>
<nav class="navbar navbar-inverse">
<div class="container">
<div class="navbar-header">
<button type="button" class="navbar-toggle collapsed" data-toggle="collapse" data-target="#navbar"
aria-expanded="false" aria-controls="navbar">
<span class="sr-only">Toggle navigation</span>
<span class="icon-bar"></span>
<span class="icon-bar"></span>
<span class="icon-bar"></span>
</button>
<a class="navbar-brand" href="#">Bootstrap theme</a>
</div>
<div id="navbar" class="navbar-collapse collapse">
<ul class="nav navbar-nav">
<li class="active"><a href="#">Home</a></li>
<li><a href="#">About</a></li>
<li><a href="#">Contact</a></li>
<li class="dropdown">
<a href="#" class="dropdown-toggle" data-toggle="dropdown" role="button" aria-haspopup="true"
aria-expanded="false">Dropdown <span class="caret"></span></a>
<ul class="dropdown-menu">
<li><a href="#">Action</a></li>
<li><a href="#">Another action</a></li>
<li><a href="#">Something else here</a></li>
<li role="separator" class="divider"></li>
<li class="dropdown-header">Nav header</li>
<li><a href="#">Separated link</a></li>
<li><a href="#">One more separated link</a></li>
</ul>
</li>
</ul>
</div><!--/.nav-collapse -->
</div>
</nav>
<header>
<h1>Bayesian Theorem</h1>
</header>
<main>
<section id="fundamentals">
<h2>Fundamentals of Fermi Questioning</h2>
<p>
Fermi questions are a type of question that can be used to test the validity of a hypothesis. The
questions
are named after physicist Enrico Fermi, who used them to test the validity of his theories. The
questions
are designed to be answered by a simple "yes" or "no" and are intended to be used in a group setting.
The
questions are as follows:
</p>
<h2>Prompt session</h2>
<p>
ask me a fermi question related to business analysis
How many businesses in the world currently use data analysis to inform their decision making?
Derrian Chia
Assuming that there are 1 billion people in the world and that 10 percent of the population is entrepreneurial in nature. that would make it 100 million businesses that are incorporated in the world. however it seems that not all businesses are digitally inclined and are adverse towards data analysis. lets put it at another 10 percent of people that use data analysis. this brings it down to 10 million businesses.
That's an interesting observation. Can you estimate the percentage of businesses that are successful in using data analysis to inform their decision-making processes?
Derrian Chia
suppose that 10 million businesses are using data analytics. but not all of them are very adept at using analytics. probably 5 percent are really driven to use analytics to make day to day observations. that brings it down to 500 thousand businesses
That's an interesting observation. Let me ask you a question related to this: If there are 500 thousand businesses that are really driven to use analytics, what might be some challenges they face in using data effectively to inform their decisions?
Derrian Chia
i would say that since there are 500 thousand businesses these would be hard core , early adopter who have a data first mindset towards data analytics. i believe they would face scrutiny from the fellow businesses who are not an early adopter. they might get discriminated against to do business with them. That would make the social argument of what challenge they are facing. Economically, building data is not cheap. Data collection is a hassle, to be able to set pipelines to have clear and accurate data which are clean is a really difficult process for businesses. many a times resources has to be allocated in order to clean data up to make sense of it.
Interesting points. Based on your analysis, it seems that there are both social and economic challenges that businesses face when trying to adopt and use data analytics. Could you elaborate further on the economic challenges businesses might face in building and maintaining a data analytics infrastructure? And what are some potential solutions to these challenges?
Derrian Chia
I would think that businesses face the economic costs of maintaining a data analytics from a macro perspective. this may include hiring costs to engage data cleaners to cleanse the data. additionally there has to be resources allocated to have people make sense of the data once collected. this may be data scientist or engineers that are hired. This would form the human capital side of things. apart from that , physical infrastructure may also be important. such as servers or gpus in order for a company to really house datas. they may have to engage a data center to allocate compute resource to this businesses to run effectively. but ultimately i believe that a good solution to this problem would be to drive the marginal cost of compute and the democratisation of data knowledge. firstly lowering the cost of compute to near zero would increase accessibility so that it becomes a basic need that everyone can gain access. democratisation of data knowledge would reduce the salaries of these data professionals. however it would benefit organisations as they have a larger pool of options to choose from and reduce their cost
That's a great analysis. Now, let me ask you another Fermi question related to this topic: How might the use of data analytics affect job security for employees in traditional business roles?
Derrian Chia
i want you to act as a fermi question coach review my fermi answers and how i can improve on them.
Sure, I'd be happy to help you improve your Fermi answers. Please provide me with your previous Fermi answers so I can review them and provide feedback.
Derrian Chia
How many businesses in the world currently use data analysis to inform their decision making?
Derrian Chia
Assuming that there are 1 billion people in the world and that 10 percent of the population is entrepreneurial in nature. that would make it 100 million businesses that are incorporated in the world. however it seems that not all businesses are digitally inclined and are adverse towards data analysis. lets put it at another 10 percent of people that use data analysis. this brings it down to 10 million businesses.
That's an interesting observation. Can you estimate the percentage of businesses that are successful in using data analysis to inform their decision-making processes?
Derrian Chia
suppose that 10 million businesses are using data analytics. but not all of them are very adept at using analytics. probably 5 percent are really driven to use analytics to make day to day observations. that brings it down to 500 thousand businesses
That's an interesting observation. Let me ask you a question related to this: If there are 500 thousand businesses that are really driven to use analytics, what might be some challenges they face in using data effectively to inform their decisions?
Derrian Chia
i would say that since there are 500 thousand businesses these would be hard core , early adopter who have a data first mindset towards data analytics. i believe they would face scrutiny from the fellow businesses who are not an early adopter. they might get discriminated against to do business with them. That would make the social argument of what challenge they are facing. Economically, building data is not cheap. Data collection is a hassle, to be able to set pipelines to have clear and accurate data which are clean is a really difficult process for businesses. many a times resources has to be allocated in order to clean data up to make sense of it.
Interesting points. Based on your analysis, it seems that there are both social and economic challenges that businesses face when trying to adopt and use data analytics. Could you elaborate further on the economic challenges businesses might face in building and maintaining a data analytics infrastructure? And what are some potential solutions to these challenges?
Derrian Chia
I would think that businesses face the economic costs of maintaining a data analytics from a macro perspective. this may include hiring costs to engage data cleaners to cleanse the data. additionally there has to be resources allocated to have people make sense of the data once collected. this may be data scientist or engineers that are hired. This would form the human capital side of things. apart from that , physical infrastructure may also be important. such as servers or gpus in order for a company to really house datas. they may have to engage a data center to allocate compute resource to this businesses to run effectively. but ultimately i believe that a good solution to this problem would be to drive the marginal cost of compute and the democratisation of data knowledge. firstly lowering the cost of compute to near zero would increase accessibility so that it becomes a basic need that everyone can gain access. democratisation of data knowledge would reduce the salaries of these data professionals. however it would benefit organisations as they have a larger pool of options to choose from and reduce their cost
Great job on your Fermi answers! Here are a few suggestions for improvement:
When estimating the number of businesses using data analysis, consider gathering data from industry reports or surveys to support your estimate.
When estimating the percentage of businesses successfully using data analysis, consider defining what success means in this context and how it can be measured.
When discussing the challenges businesses face in using data effectively, consider providing specific examples of businesses that have faced these challenges and how they overcame them.
When proposing solutions to these challenges, consider discussing potential drawbacks or limitations of these solutions and how they can be addressed.
Overall, keep in mind that Fermi questions require creative thinking and estimation based on available information, so don't be afraid to take risks and make assumptions in your answers.
Derrian Chia
what is the point of fermi questioning
The point of Fermi questioning is to encourage people to think critically and creatively about complex problems or questions that may not have an easy or straightforward answer. It involves breaking down a problem or question into smaller parts, estimating or making educated guesses about various aspects of the problem, and then using those estimates to arrive at a reasonable answer or solution. The process encourages curiosity, open-mindedness, and flexibility in thinking, and can be useful in a wide range of contexts, including science, business, and everyday life.
Derrian Chia
why would someone want to improve on fermi questioning
There are several reasons why someone might want to improve on their Fermi questioning skills.
Firstly, being able to ask thoughtful and insightful Fermi questions can help to develop critical thinking skills. By asking questions that require estimations and logical reasoning, one can improve their ability to analyze and evaluate information.
Secondly, Fermi questions can be a useful tool in problem-solving and decision-making. They can help to break down complex problems into smaller, more manageable parts and can provide a framework for approaching solutions.
Finally, Fermi questioning can be a useful communication tool. By asking well-crafted questions, one can engage in meaningful discussions and debates, clarify misunderstandings, and promote understanding and learning.
<div>
</div>
</p>
</section>
<section id="history">
<h2>History of Bayesian Theorem</h2>
<p>
Bayesian theorem is named after Reverend Thomas Bayes, who first formulated the theorem in the 18th
century. He
developed the theorem in the context of solving a specific problem in probability theory known as the
"inverse
probability problem." However, it was not until after Bayes' death that his work was published by
Richard Price,
another prominent mathematician of the time.
</p>
<p>
In the years that followed, Bayesian theorem gained wider acceptance and was further developed by
mathematicians
such as Pierre-Simon Laplace. Over time, the theorem became an essential tool in the field of statistics
and
probability, with numerous applications across various disciplines, including science, engineering,
economics, and
more recently, artificial intelligence and machine learning.
</p>
</section>
<section id="application">
<h2>How Bayesian Theorem Applies to Large Language Models</h2>
<p>
Bayesian theorem is an essential component of many machine learning algorithms, including large language
models
(LLMs). In the context of LLMs, Bayesian methods can be used to estimate the parameters of the model,
update
beliefs about the model's parameters as new data is observed, and make predictions based on the current
state of
the model.
</p>
<p>
One common application of Bayesian theorem in LLMs is through the use of Bayesian inference, where the
model's
parameters are updated iteratively as new data is processed. This allows the model to learn from the
data and
make more accurate predictions over time.
</p>
</section>
<section id="effectiveness">
<h2>Why Bayesian Theorem Works So Well in Large Language Models</h2>
<p>
Bayesian theorem works well in large language models because it provides a principled framework for
updating
</p>
</section>
<section id="application">
<h2>How Bayesian Theorem Applies to Large Language Models</h2>
<p>
Bayesian theorem is an essential component of many machine learning algorithms, including large language
models
(LLMs). In the context of LLMs, Bayesian methods can be used to estimate the parameters of the model,
update
beliefs about the model's parameters as new data is observed, and make predictions based on the current
state of
the model.
</p>
<p>
One common application of Bayesian theorem in LLMs is through the use of Bayesian inference, where the
model's
parameters
</p>
</section>
<section id="backpropagation">
<h2>How does back propogate and Bayesian Theorem Works Together</h2>
<p>
Bayesian theorem and backpropagation are two different ideas, but they can work together in some
situations.
Imagine that Bayesian theorem is like a detective who uses clues to guess who might be the criminal.
Backpropagation is like a teacher who helps a student learn by telling them which answers are right and
wrong.
When we have a big computer brain called a neural network, we sometimes want to help it learn using both
the detective's clues (Bayesian theorem) and the teacher's advice (backpropagation). In some cases, they
can work together to make the computer brain smarter and better at understanding things.
</p>
<!--- what you have assumed -->
<h1>
Reflections
</h1>
<p>I am assuming that P(A) is probability of taking the coin of A.</p>
<p> I am then assuming that P(B) is the proability of taking coin B</p>
<p> i have no idea what is the calculation for posterior distribution> seems like it is a
distribution, means a bunch of numbers or graph. it needs to be based of existing knowledge or smt
</p>
<!-- # what you have understood -->
<h1>
What I have understood
</h1>
<p>The theorem is written as P(A|B) = (P(B|A) * P(A)) / P(B), where P(A|B) is the probability of event A
happening,
given that event B has occurred, P(B|A) is the probability of event B happening, given that event A has
occurred, and P(A) and P(B) are the probabilities of events A and B happening independently.
</p>
<p>
i need to ignore the conditional probability first because i dont know what is that value.
</p>
<p>
coin a is 0.5 and coin b is 0.75. but given the coin tosses, it is telling us whether it is heads or tails
</p>
<p>
this means that i have P(T) is actually 5 tosses, 0.25 and P(H) is actually 0.75 probability which is 15 times.
total times is actually 20 times.
instead of assuming of which coin did i take.
i want to assume which probability of it is heads and which probability is tails
P(T)/P(H) = 0.25/0.75 = 0.3333333333333333
but what is the conditional probability of B given A?
P(B|A) = ?
P(B|A) = 0.25*0.75 = 0.1875
then what about p(A|B)?
= 0.1875*0.3333333333333333 = 0.0625
so the conditional probability of A given B is 0.0625
So what can i do with the conditional probability?
</p>
<h1> what went wrong </h1>
<p>
<ul>
<li>
0.5 assumption and 0.75 assumption is wrong.
</li>
<li>
i used the fraction of the probability to sub into the equation
</li>
<li>
instantly assuming that the probability is how much.
</li>
<li>
what helped is and using actual values to solve the equation.
</li>
<li>
another misconception is that the distribution is a single number. NO the distribution is a bunch of number. it distributes
</li>
</ul>
</p>