Kaggle DNN Project
Develop an algorithm or learning system that can solve each task, separateley,
without human interaction. This can be acheived through the use of a single
learner, an ensable of multiple learners, architecture search, curriculum learn-
ing, or any other technique, as long as task-specific model parameters are not
human-defined. Choose two of the tasks provided, use the same method
(architecture / algorithm) in both tasks and investigate the generalization
of your algorithm.
1.3 I’m Something of a Painter Myself
A GAN consists of at least two neural networks: a generator model and a dis-
criminator model. The generator is a neural network that creates the images.
In this competition, you should generate images in the style of Monet. This
generator is trained using a discriminator. The two models will work against
each other, with the generator trying to trick the discriminator, and the dis-
criminator trying to accurately classify the real vs. generated images. Your task
is to build a GAN that generates 7,000 to 10,000 Monet-style images.
2 The Report
2.1 Abstract
One paragraph with the major aspects of the entire project in a prescribed
sequence that includes:
1. The overall purpose of the study and the research problem(s) you investi-
gated
2. The basic design of the study
3. Major findings or trends found as a result of your analysis
4. A brief summary of your interpretations and conclusions
2.2 Introduction and Objective
Give a clear idea of the challenge you choose to address. Provide some back-
ground information on the specific problem or issue you are addressing, explain
the objective and clearly outline your proposed solution.
2.3 Data Description
Elaborate all information about the data and the ground truth in a way that
allows someone else to use the data in a future research project and understand
the data’s content and structure. In this section you should provide Exploratory
Data Analysis (EDA), i.e., visually present the data, and explore it.
22.4 Methods and Algorithms
Present the methods and algorithms you use, explain about each step (i.e the
chosen augmentations/ model/ loss function etc.). Make sure to present the ar-
chitecture, loss functions, learning paradigm and hyper-parameters, if you use
augmentation discuss it and present examples.
Explain the motivation behind your choices and reference existing work.
When you approach this problem remember that you have many things you
have learned in your arsenal such as: augmentation, drop outs, regularization,
transfer learning, test time augmentation and many more.
2.5 Challenges and Difficulties
Detail as much as possible about the problems you faced and how you over-
came them (such as challenges from the data or from the task at hand), the
assumptions you made in the project and why you made them.
2.6 Experiments
Presentation of the experiments, what hypotheses have you examined and how.
Explain about objective and subjective metrics that you used and why did you
choose them. Show the significance of each component in your method using
ablation study. Present plots of the epochs with respect to training/validation
loss analyze the plots; use semantics such as overfitting/underfitting. Present
your results both qualitatively and quantitavely . Present the measures that
you use.
2.7 Results
Objectively present your key results, without interpretation, in an orderly and
logical sequence using both text and illustrative materials (Tables and Figures).
Refer to figures and tables that present your results.
Add your results from the competitions leader-board.
* In the experiments and results paragraphs, you should refer to topics taught
in the course such as overfitting and regularization.
2.8 Conclusions and Summary
Conclude and summarize your overall results and major findings.
2.9 References
List of the references that you actually cited in the body of your report.
3
Buy an Essay Online from Professional Assignment Writing Agency. RoyalResearchService.com work on all types of assignments -irrespective of their academic field and level of difficulty. We handle small and medium sized papers, a few pages long essays, as well as full-scale dissertations/theses and coursework. All our papers are written from scratch.
Need Help Writing an Essay?
Tell us about your assignment and we will find the best writer for your paper.
Write My Essay For MeNeed Assignment Writing Help? We write papers from scratch and within your preferred time frame. The simple, safe, and Cheap way to buy Original assignment answers.
