Blog

If you want to understand the working of any complex system then you would first try to break everything down to its fundamentals. Then you would understand the fundamentals very clearly and then use that understanding of the fundamentals to completely understand the system. This is the very basis of Physics, which is one of the most successful endeavours of humankind. We have been able to take this approach whereby we break down systems to fundamentals and then use its knowledge to have the most detailed and clear understanding of lot of things.

"ऊ ठ्याक्कै सारापोभा जस्तै छे।"

In the [previous post][fnn] we discussed how to make functional Neural Network (NN) with julia. In [another post][nn-package] we discussed how to package our NN into an independent julia package. As described in [that post][nn-package], we are able to import the new package as

In a seris of previous posts we, described how to make a [fully functional Neural Network][fnn] in julia. Lets create a package so that we can import the neural network in julia as a regular package available in every session.

Say you wrote an ingeneous julia function

As we saw in [previous post][multinode] addition of any number of layers and any number of nodes will only be able to model a linear functions. If we want our network to also be able to learn non linear functions, then we should introduce some non linearlity to the network. One way of adding non linearity is to use so called _activation function_. The idea is that we pass the output of each layer through a (non linear) function. Our examples so far can be considered to have been using a linear activation function. In particular, they all used identity activation function

After a series of [previous posts][activation-func] we hav learned the key steps of making a Machine Learning:

In the previous two posts we have already seen how to construct a single neuron neural network with or without hidden layer. If we want to input multidimensional input and get multidimensional output from the netwro then we have to add more nodes in each layers so that each node in the layer can process independent multidimensional input. For this purpose in this post we will construct a network with a single output layer but with multiple nodes in input layer. This is useful for example if we want to train our network to learn the scalar (one number) function of $N$ dimensional vector.

This is a content.

Similar to [last post](/posts/2022/05/basic-neural-network-with-julia/), where we made a basic neural network with just one neuron and no hidden layer, we will build a two neuron and a single hidden layer.

In this post I will try to describe building a basic regression model with gradient descent. This forms a very basic prototype of neural network. This by no means is a full fledged neural network, but this makes very fundamental foundation of neural network. Feedforward to calculate the error and back propagation to update the weights and biases from the error function. The code cells can be collected into a script and run as an individual julia script to reproduce everything that is described here in this article.

केपी ओलीलाई मैले बृहत्तर रूपमा एक चटकेका रूपमा बुझेको थिएँ त्यसमा मेलै पुनर्विचार गरेको छु।

My idea of Mathematics was of this beautiful tool that worked so consistantly. Something that fascinated me with mathematics in my early days of introduction of it with addition and subtraction is that there were different approaches to solving problem (adding and and subtracting which I much later now know as commutative and associative property of an additive group) and yet they always gave same result. Sometimes I secretly wished that it didn't work, I hoped that somehow we came to inconsistant result, but it was never so. Well actually at around 6th grade in school, when we first learned algebra, I knew of the bogus proof that 1:2, and there were various variants of it. This gave me a glimmer of hope that there might be inconsistancies in mathematics afterall. I later knew that each variant of this proof explotied one way or the other the use of cancellation of zeros in each side of equation, which is not allowed.[Read more on this].

गीट भनेको फाइलको भर्सन मिलाउने एउटा अनुप्रयोग हो। कुनै पनि परियोजनामा कुनै फाइलहरूको विभिन्न भर्सन राख्नु पर्छ भने यो प्रयोग गर्दा धेरै सजिलो हुन जान्छ।

Neutrinos are little tiny particles that carry no charge and very little mass. The exact mass of neutrino is not known yet but, we know that they have very tiny mass. The confirmation of neutrino oscillation confirms that at least one neutrino is massive. In the current theory of neutrinos, we have three flavors of neutrinos corresponding to each charged lepton flavors.

भाषा साहित्य र गणितका सुत्रहरूका बीचको समानताको व्याख्या।

How I had to see two dying fish in front of my eyes and couldn't do anything about it

calculation of probabilities in the game of dice wars

मनुष्यका केही समस्याहरूबारे मेरो छोटो टिप्पणी

नेपाली शब्दान्तका केही स्वर वर्णका उच्चारणमा मैले देखेको केही प्रवृत्तिहरूका बारेमा मेरो ब्यक्तिगत टिप्पणी

रत्ती आँशु तिमी नरोइदेऊ \\

Messi has just won the sixth Ballon d'Or. I discuss briefly about my take on this and messi in general.

very brief scribble on a thought about insanity

अनिर्णयको शिकार भएर मानिस कसरी यथास्थितिको बन्दी बन्न पुग्छ, अनि यसका परिणामहरू के हुन्छन् भन्ने बारे

नेपाली भाषाको पछिल्लो स्थिति यसका करणहरू र समस्याहरूका समाधनको केही बुँदागत प्रयासहरूबारे मेरो टिप्पणी

a stub possible model for free will.

कम्प्यूटरमा नेपलाी भाषा लेख्ने पुरानो र नयाँ यूनिकोड तरिकाबारे छोटो विश्लेषण।

very brief introduction to the Dirac equation

a brief look into Einstein's field equation

a brief look into QED and its working in under half an hour.

an introduction to the Dirac equation

प्रमाणित भयो आज

a short foray into spherical pendum solution with lagrangian done with sympy.

यामले हेमन्त्त हो यो। बाहिर बर्षको पहिलो हिँउँ बाक्लो परेको छ, सेताम्म्य बाहिरको दृष्य हेरिरहँदा मनमा चिसो पस्यो। तापक्रमको हिसावले, मेरो कोठा, जहाँ बसेर म बाहिर नियालिरहेको छु, १८ देखि २० सेल्सियसको हाराहारीमा होला। तर मनमा पसेको चिसो तापक्रमसँग सम्बन्धित थिएन। बाहिर निकैबेर नियाल्न मन लाग्यो, मेरा नजरहरू फुरूरूरू उडेका हिउँका पत्ताहरूमा भएका भएपनि वातावरणमा घटेका घटनाहरूसँग खासै मेरो दिमागी एकाग्रता थिएन, दृष्यहरू आँखाबाट पसेर दिमागमा नपुगी बीचै बिलाइरहेका थिए। सोचको डोरो अन्यत्रै ब्यस्त थियो। यसैबीच दृष्यमा मनुवाहरू समान्य रूपमा आवतजावत गरिरहेकै थिए। छाला खासै चिसो नभएपनि दिमाग झन् चिसो हुँदै थियो।


एकछिन आँखा चिम्म गरेर गम खाने कोशिस गरेँ। आफ्नो सन्तानको निधनले आमा सीता दाहाललाई जत्तिको पीडा अरू कसैलाई भएन होला। बाबु प्रचण्डलाई आफ्नो प्रमुख राजनीतिक सहयोगी त्यसमाथि आफ्नै छोरा गुमाएकोमा ठूलो पीडा भयो होला।

सबैकुरा पृथ्वीतिर तानिने कुरालाई ब्याख्या गर्न जुन समिकरणको प्रयोग गरिन्छ त्यही समिकरण प्रयोग गर्दा भीमकाय ब्रह्माण्डीय पीण्डहरूको ब्यहोरा लाई पनि सजिलै वर्णन गर्न सकिन्छ। चन्द्रमाले पृथ्वीलाई किन घुम्छ, ग्रहण किन र कसरी लाग्छ भन्ने कुरा त्यही सामान्य सिद्धान्तका आधारमा ब्याख्या गर्न सकिन्छ। कुनै कुराले हामीतर्फ प्रकाश दिँदा (वा परावर्तन गर्दा) हामीले तिनलाई देख्न सक्छौँ। त्यसै सिद्धान्तका आधारमा हामीले आकाशमा ताराहरूलाई देख्न सक्छौँ किनकी ताराहरूले प्रकाश उत्सर्जन गर्छन् । परावर्तन गर्ने प्रकाशका आधारमा हामीलाई हाम्रो पृथ्वीको एउटा चन्द्रमा छ भन्ने थाह छ। उत्सर्जन गर्ने आधारमा सूर्य कहाँ र कतिवाट छ भन्ने थाहा छ।कति बलले कुनै वस्तुलाई तान्छ भन्ने आधारमा हामीले वस्तुको पीण्ड कति छ भन्ने कुरा पत्ता लगाउँछौ। यिनले अरू पीण्डलाई आकर्षण गरेको आधारमा हामीले तिनको पिण्ड र सङ्ख्या भन्न सक्छौँ। हेरेर (यसले उत्सर्ग गर्ने वा परावर्तन गर्ने प्रकाशका आधारमा) हामीले ती कति सङ्ख्यामा वा कहाँ कहाँ छन् भनेर पत्ता लगाउन सक्छौँ। र यी दुवै तरिकाबाट हामीले पत्ता लगाएका ग्रह वा तराको सङ्ख्याका बीच मेल खाएको छ।कुरा यत्ति हो।

a very short look of discrete Fourier transform of Gaussian function

A basic exercise on calculating the trajectory of charged particle in cross field

A very basic problem of trajectory of particle with under linear drag force F= -k v

An implementation of Huffman algorithm in python programming language

एक दिन क्यामपसबाट फर्कँदाको अनुभव सम्बन्धी छोटो अनुभव

A brief descussion on whether randomness in nature is fundamental.

my rant to myself about an unproductive self observant day inside robotics club Pulchowk Campus, Tribhuvan University

I was sitting on my chair in the room. Because of a stream of thinking I got lost from the world outside. But when I got out of the imagination and landed back on my chair I suddenly realized everything around me has expanded proportionally i.e. they look bigger than I was used to seeing them. I couldn’t compare them to outside world; but for me everything inside my room looked expanded.

memoir of few observation I made during my commute from Tri-Chandra campus to Pulchowk Campus

My thoughts on the flow of time and its finite nature, and the possible and eventual end of everything

What would it mean if I had to die early before really I could do something? I’ve dreams fantasies and promises to reach but if it were such that I had to leave all these beautiful things around so easily and never be able to come back. Each days now I have dreamt and promised more and more and not done anything at all. The promises fantasies and dreams are piling up; I see beauty in doing them; an art of being alive. If I were to really go without even being able to realize those fantasies, would that be tolerable?

This article I wrot ebasically came out of my discussion with my friend Sagar Humagain. I actually wrote in this from during class 11, as a speech during the morning assembly which I never delivered.

Every things start to work on your favor. Every matter which seemed somehow to have lost track seems to come back to track and everything, which denied going on your favor, starts following you.