Sergiy Isakov
Private message
Offer work
Kharkiv,
Ukraine
available for hire
Rating
Skills and abilities
Programming
Services
Photo, Audio & Video
Portfolio
-
The artificial immune system.
AI & Machine Learning
Use of the principles and algorithms of artificial immune systems in creating a neuron network to classify objects with small training data sets. The main idea is to clone the vector of the internal layer of the NS to improve the efficiency of the classification process.
-
GAN for Images Super Resolution
AI & Machine Learning
Development of ML models based on Generative Adversarial Network (GAN),
which enhances the content loss function with competitive loss by training GAN.
The model enables the reconstruction of images with a large scal rate (in 4 times) and was used to reconstruct the medical images of the microscope.
-
Not to run on Pytorch for RETS
AI & Machine Learning
Development of a model for the price forecast of real estate on the basis of Artificial Neural Networks (ANN) in Pytorch
The first step is the preliminary processing of the source data
Data from the MySQL database (access to the server is provided by the customer)
- data cleaning, emissions analysis, filling missed data, etc.
… Analysis of interconnection, interaction and correlation of data. Selection of independent parameters
The second phase is the restoration of sales prices.
The development of model
Training, testing and so on.
- "rehabilitation" of the sales price for objects for which the sales value is not indicated in the database (this objects 400)
- verifying the accuracy of the model compared to the real price of the sale of such objects
The third phase is the forecast of the sales price for 1-5 years:
Classification of source data
Formation of temporary dependencies
- analysis of stability, trends, timelines, periodicity, abnormal areas, etc.
Building a model for prediction
-
The Neuronal Network of Robot Movement
AI & Machine Learning
Robot movement control is built on the basis of Artificial Neural Networks (ANN)
ANN has two nodes (neuron) on the input layer, ten nodes on the hidden layer and seven nodes on the output layer.
The nodes on the input layer correspond to the data (signal levels) of the left and right on the course of the movement of the robot sensors
Signals from the sensors are normalized by the maximum level of the signal from the sensor in such a way that the input layer nodes are given values in the range of 0 to 1.With several sensors on each side, their indications are summarized with different weight coefficients and submitted to the input nodes as a aggregated signal.
… The connections between the entrance nodes and the nodes of the hidden layer are built on the basis of the relay function, and between the nodes of the hidden layer and the output - on the basis of the sigma function
The result of the forecast of action is formed in seven nodes of the output layer and ultimately formed from the condition of the highest probability, which corresponds to one of the seven categories (action), which corresponds to the following:
“Keep move – both rotate in the same direction”
Action = 1 - 'Right brake, left rotate - turn right'
Action = 2 - 'Right reverse, left rotate - cool right'
Action = 3 - 'Left brake, right rotate - turn left'
Action = 4 - 'Left reverse, right rotate - cool left'
Action = 5 - 'Both brake - smooth braking'
Action = 6 - 'Both reverse - emergency braking'
-
Analysis of energy distribution
AI & Machine Learning
preliminary data processing of sensors, model development, forecasting, visualization
-
Classification of Users
AI & Machine Learning
preliminary processing of user data, development of user classification model
-
Prognosis of sales
AI & Machine Learning
Prognosis of sales on the basis of timelines
Reviews and compliments on completed projects 10
-
Real expert -
Craft master -
Quick answers -
First-class quality -
Nice communication -
High responsibility -
Great price -
Lightning fast
25 April 2023
412 USD
Improvement of the previous ML project
Sergey launched on a server created in the previous ML project.
Qualified I went to all stages, everything was done professionally, thank you for cooperation.
I recommend it!
29 March 2023
826 USD
Machine Learning для підбору рівнів в грі вікторині
Sergey was very responsible for the task. Before creating the model, he familiarized with the product in which it will be used and proposed the solution that will best fit in this task. Communication was without problems.
I recommend cooperation!
3 February 2023
180 USD
Prediction and validation of the value of the goods through ML.
Все отлично, работа выполнена быстро и качественно!
25 January 2023
338 USD
Development of a neuron network for the classification of acoustic signals
The commissioner showed a high level of competence and performed the work earlier than the deadline. I recommend !
|
3 May 2022
226 USD
Проверка алгоритма
Sergey demonstrated excellent work in time, a professional approach with the addition of creative ideas that in the essence of the topic.
22 April 2021
250 USD
Обработка данных при помощи LSTM
Спасибо за качественно и в срок выполненную работу! Исполнитель продемонстрировал готовность работать в новой области и продемонстрировал высокий профессионализм.
13 December 2020
45 USD
Data analysis in Python – regular collaboration
Both communication and work - everything is fast, careful and in essence, quality. I am satisfied. Thank you !
22 June 2020
81 USD
Разработка CNN модели (на Python) для распознoвания тональности текста
рекомендую, было приятно работать!
18 March 2020
131 USD
CNN + AIS
Была проведена хорошая работа. Фрилансер проявил себя ответственно и даже творчески (там где это было нужно). Работой доволен. Буду рад будущему сотрудничеству.
24 January 2020
27 USD
Pytorch for RETS deep learning
Thank you for the successfully completed project!
Activity
| Latest proposals 10 | Budget | Added | Deadlines | Proposal | |
|---|---|---|---|---|---|
|
Freelance project
226 USD
|
|||||
|
Freelance project |
|||||
|
Freelance project
564 USD
|
|||||
|
Freelance project |
|||||
|
Stabilize Image Processing Pipeline (DL + CV) for Tomography / Ophthalmic Imaging
3000 USD
|
|||||
|
Connecting AI to the database
|
|||||
|
AI - select and configure the RVC model
226 USD
|
|||||
|
Voice request to search for real estate listings in the database
|
|||||
|
Laboratory of Machine Learning
21 USD
|
|||||
|
Machine learning lab
21 USD
|