Sökning: "Backpropagation through time"

Hittade 5 uppsatser innehållade orden Backpropagation through time.

1. 1. Adversarial robustness of STDP-trained spiking neural networks

   Kandidat-uppsats, KTH/Skolan för elektroteknik och datavetenskap (EECS)

   Författare :Karl Lindblad; Axel Nilsson; [2023]
   Nyckelord :;

   Sammanfattning : Adversarial attacks on machine learning models are designed to elicit the wrong behavior from the model. One such attack on image classifiers are maliciously crafted inputs that, to the human eye, look untampered with but have been carefully altered to cause misclassification. LÄS MER

3. 2. Explaining Neural Networks used for PIM Cancellation

   Master-uppsats, KTH/Skolan för elektroteknik och datavetenskap (EECS)

   Författare :Fredrik Diffner; [2022]
   Nyckelord :Telecommunication; Passive intermodulation; PIM cancellation; Fully connected neural network; Convolutional neural network; Time series regression; Explainable AI; backpropagation-based XAI methods; perturbation-based XAI methods; Telekommunikation; Passiv intermodulation; PIM-eliminering; Fullt sammankopplade neuronnät; Konvolutionerande neuronnät; Tidsserieregression; Förklarande AI; Bakåtpropagerande-baserade XAI-metoder; Perturbations-baserade XAI-metoder;

   Sammanfattning : Passive Intermodulation is a type of distortion affecting the sensitive receiving signals in a cellular network, which is a growing problem in the telecommunication field. One way to mitigate this problem is through Passive Intermodulation Cancellation, where the predicted noise in a signal is modeled with polynomials. LÄS MER

4. 3. Forecasting the USD/SEK exchange rate using deep neural networks

   Kandidat-uppsats, Lunds universitet/Matematisk statistik

   Författare :Thomas Hamfelt; [2019]
   Nyckelord :Neural networks; Deep Learning; Foreign Exchange; Finance; Mathematics and Statistics;

   Sammanfattning : This thesis is about predicting the average ten minute closing bid price of the USD/SEK exchange rate by applying deep learning methods. First, the time lag method is applied for the vanilla Feedforward Neural Network (FNN) to undertake one-step prediction. LÄS MER

5. 4. Real-time Process Modelling Based on Big Data Stream Learning

   Master-uppsats, Mälardalens högskola/Akademin för innovation, design och teknik

   Författare :Fan He; [2017]
   Nyckelord :control system; real-time process; deep learning; recurrent neural network; Backpropagation through time; Kalman filter;

   Sammanfattning : Most control systems now are assumed to be unchangeable, but this is an ideal situation. In real applications, they are often accompanied with many changes. Some of changes are from environment changes, and some are system requirements. So, the goal of thesis is to model a dynamic adaptive real-time control system process with big data stream. LÄS MER

7. 5. Autonomous Calibration and Control of Mine Vehicles

   Master-uppsats, Institutionen för datavetenskap; Tekniska högskolan

   Författare :Malcolm Lerider; [2013]
   Nyckelord :atlas copco; autonomous; calibration; mine vehicles; lhd; steering;

   Sammanfattning : The mining industry desires to cut costs and to operate in more dangerous mines, which is why companies such as Atlas Copco are developing autonomous vehicles. The problem to navigate autonomously is however complex, so the vehicles have in the recent years become more intelligent; the number of computers, actuators and sensors are increasing. LÄS MER