computerphysicslab

"Convolution is probably the most important concept in deep learning  ... You can imagine convolution as the mixing of information..."

"There can be a lot of distracting information in images that is not relevant to what we are trying to achieve ... the shape of a blouse is very different from the shape of a shirt, jacket, or trouser. So if we could filter the unnecessary information out of images then our algorithm will not be distracted by the unnecessary details like color and branded emblems. We can achieve this easily by convoluting images with kernels."

"Instead of having fixed numbers in our kernel, we assign parameters to these kernels which will be trained on the data. As we train our convolutional net, the kernel will get better and better at filtering a given image (or a given feature map) for relevant information. This process is automatic and is called feature learning."

"A quantum algorithm would be able to calculate all possible combinations described by the kernel with one computation and i

computerphysicslab

"Latent Dirichlet Allocation (LDA) Blei et.al, 2002, proposed LDA model. They described thismodel as a generative probabilistic three-level hierarchical Bayesian model. They say that this model is for large collections of discrete data and tries to find short descriptions for the collection to process large collection of documents. It models the corpus as a collection of documents, each document as a distribution of multiple topics and each topic as a mixture of words."

"rstenjak et.al, 2013, proposed a possibility of using a KNN with TF-IDF method for text classification. The evaluation is based on the speed, accuracy and quality of classification. The results include both good and bad features. The main motivation for this paper was to develop concept frameworks with emphasis on KNN & TF-IDF module."

https://www.digitalxplore.org/up_proc/pdf/268-148653117584-89.pdf

#knn #lda #topicModel #textClassification

computerphysicslab

Sugerencias kNN

"Un conocido algoritmo de Data Science y Machine Learning es k-Nearest-Neighbor (kNN ó K-vecinos más cercanos). ARN aplica este algoritmo para generar sugerencias de perfil. De este modo, si un profesional dispone de las habilidades de PHP y jQuery, posiblemente reciba una recomendación de completar su perfil con un test de AJAX. El algoritmo no conoce de forma ad-hoc la conexión entre AJAX y PHP+jQuery, sino que simplemente analiza los perfiles más parecidos al profesional (K-vecinos más cercanos) y observa una tendencia que no está presente en el perfil del sujeto. De esta forma, se aplica inteligencia colectiva para que ARN aprenda las relaciones entre la compleja taxonomía de skills profesionales. Obviamente los skills identificados como de potencial interés para el usuario han de plantearse como sugerencias."

#kNN #kNearestNeighbors #machineLearning #inteligenciaColectiva #inteligenciaArtificial

--

KNN recommendations

computerphysicslab

Dataset: over 200,000 scholarly articles, including over 100,000 with full text, about COVID-19, SARS-CoV-2, and related coronaviruses.

https://www.kaggle.com/allen-institute-for-ai/CORD-19-research-challenge
https://www.kaggle.com/allen-institute-for-ai/CORD-19-research-challenge?select=document_parses

computerphysicslab

N-GRAM:  Department of Health and Human 144
N-GRAM:  Blue Cross and Blue Shield 75
N-GRAM:  Providing personalized assistance to customers 42
N-GRAM:  Food and Drug Administration FDA 42
N-GRAM:  Country of development US Type 42
N-GRAM:  Department of Housing and Urban 31
N-GRAM:  Clin Pract Cases Emerg Med 25
N-GRAM:  Chan School of Public Health 20
N-GRAM:  Country of development China Type 18
N-GRAM:  Severe acute respiratory syndrome coronavirus 17

computerphysicslab

37819 covid
10693 coronavirus
8501 sars
7614 virus
7415 cov
5199 emergency
4533 respiratory
4220 infection
3984 pandemic
3671 cdc

computerphysicslab

"...it remains a question of whether the model has learned to do reasoning, or simply memorizes training examples in a more intelligent way."

When GPT-3 is asked "do you really think you exist at all?"

It answers: "I think I exist, but I can't be sure"

So, defenitely GPT-3 does not do reasoning.

The most solid reasoning of all times is Descartes cogito ergo sum, not applied by GPT-3

computerphysicslab

rm go.mod 
go mod init goSNMP
	go: creating new go.mod: module goSNMP
cat go.mod 
	module goSNMP
	go 1.13

En main.go:
	import ("goSNMP/snmpPrintersLib")

computerphysicslab

RANK	LENGTH	KEYWORDS	URL	SNIPPET
--------	------	-----------------	----------------	--------------------------------------------------------------------------------------------
146.02	226	[covid pandemic health ways response research reduce]	https://www.apprise.org.au/project/nhmrc-extends-apprise-funding-for-critical-covid-19-projects/	* Generate evidence that can be applied to clinical practice during the pandemic, regarding the impact of multiple candidate interventions to reduce mortality or reduce the length of intensive care unit admission or both in critically ill patients with COVID-19 infection

computerphysicslab

covid
coronavirus
sars
pandemic
epidemiology
immunology
immunity
vaccine
hydroxychloroquine
lockdown