Scientific review

 I chose the paper Generative Adversarial Networks[1] from Ian Goodfellow, which I am going to evaluate how good the paper is. I am going to use the IJCAI review template to asses the quality of this paper.

Relevance

– Is the paper fully within the scope of the conference? Will the questions and results of the paper be of interest to researchers in the field?


This paper has demonstrated the viability of the adversarial modeling framework, suggesting that these research directions could prove useful. It opens up quite a few paths, as a lot of problems have a limited data set, so now it would be possible to enrich synthetically those data sets.

Significance

– Is this a significant advance in the state of the art? Is this a paper that people are likely to read and cite? Does the paper address an important problem? Does it open new research directions? Is it a paper that is likely to have a lasting impact?


This is a significant advance in the state of the art and shows a novel idea. As it is something completely new, that potentially is able to enrich data sets, it will be read and cited by others.
==> Score 7

Originality

– Reviewers should recognise and reward papers that propose genuinely new ideas. As a reviewer you should try to assess whether the ideas are truly new. Novel combinations, adaptations or extensions of existing ideas are also valuable


The idea of the paper is new and it clearly shows how it is different to the other state of the art approaches and where the pros and cons are of this novel approach. It is something completely new and never seen.
==> Score 9

Technical Quality

– Are the results technically sound? Are there obvious flaws in the conceptual approach? Are claims well-supported by theoretical analysis or experimental results? Are the experiments well thought out and convincing? Will it be possible for other researchers to replicate these results? Is the evaluation appropriate? Did the authors clearly assess both the strengths and weaknesses of their approach?


Every claim is supported by figures, equations or explanations. The dataset they are using for the experiment is introduced. Also the architecture of the framework in the experiment is clearly described. The final result is clearly measured and compared to other models, that wanted to solve the same problem. 
==> Score 9

Clarity and quality of writing

– Is the paper clearly written? Is there a good use of examples and figures? Is it well organized? Are there problems with style and grammar? Are there issues with typos, formatting, references, etc.? It may be better to advise the authors to revise a paper and submit to a later conference, than to accept and publish a poorly-written version. However, if the paper is likely to be accepted, please make suggestions to improve the clarity of the paper, and provide details of typos.


The logic of how they describe the framework makes sense and it follows the same principle as if you would implement it, which I quite like as a Computer Scientist. For each additional element added to the framework, a paragraph describes why it is needed and how it interacts with the others. This is super helpful, as it makes the paper much easier to read and you get the idea faster. Also this is needed as the content is quite difficult.
Furthermore, they use helpful graphics, which are very well described (partly half a page long) and they describe the algorithm with pseudocode (Latex algorithm group).
In the paper and their proposed solution they use quite a lot of mathematical expressions, formulas and equation which are well formatted using LaTex. Every abbreviation used is first introduced in the text. The sentences are rather short, they don't use any fancy words and they are quite specific in explaining it.

==> Score 9

Scholarship, i.e. scientific context

– Does the paper situate the work with respect to the state of the art? Are relevant papers cited, discussed, and compared to the presented work?


Overall Score

9- An excellent paper, a very strong accept


Confidence in my assessment

7 - I have up-to-date knowledge in the area


Comments to Authors

They summarize key components, as training, evaluation, model design and sampling over other state of the art models they have used in the paper in a table with keywords and point out challenges for each model and compared it to their novel framework, which I realy like.

Sources

[1] Ian J. Goodfellow. Generative Adversarial Networks.https://arxiv.org/abs/1406.2661. Accessed: 2021-04-23

Comments

  1. Danny Cooper8 February 2022 at 03:07

    Very Informative and creative contents. This concept is a good way to enhance the knowledge. thanks for sharing. Continue to share your knowledge through articles like these, and keep posting more blogs. visit below for

    Data Engineering Solutions 

    AI & ML Service

    Data Analytics Solutions

    Data Modernization Solutions

    ReplyDelete
  2. Afrooz7 June 2022 at 17:03

    Dear Antonio

    Thank you for this interesting review. I really liked how you stated what questions each part answers. It really helps the readers understand what they should expect in each part.

    All the best,
    Afrooz

    ReplyDelete

Post a Comment

Popular posts from this blog

Blog post 3: Understanding content

Image
I chose the paper Generative Adversarial Networks[1] from Ian Goodfellow, which I am going to review what the main points are and how it works. Intorduction Generative Adversarial Networks, or GANs for short, are an approach to generative modeling using deep learning methods. Generative modeling is an unsupervised learning task in machine learning that involves automatically discovering and learning the regularities or patterns in input data in such a way that the model can be used to generate or output new examples that plausibly could have been drawn from the original dataset.[1,2,3,4] How does it work? GANs consists of two networks, a Generator G(x), and a Discriminator D(x). They both play an adversarial game where the generator tries to fool the discriminator by generating data similar to those in the training set. The Discriminator tries not to be fooled by identifying fake data from real data. They both work simultaneously to learn and train complex data like audio, video or ima
Read more
 Blog Post 2: Writing Style This second post is describing, what makes a writing output a good (or bad) piece of text, especially regarding the organization of the paper, wording without looking at the content.  I chose the paper Generative Adversarial Networks from Ian J. Goodfellow (https://arxiv.org/pdf/1406.2661.pdf), as I am very interested in this topic, my master thesis is about GANs and this paper is already familiar with me. It propose a new framework for estimating generative models via an adversarial process, in which the researchers simultaneously train two models: a generative model G that captures the data distribution, and a discriminative model D that estimates the probability that a sample came from the training data rather than G. This paper is often cited and is also the starting point for anyone who learns about GANs, even though that paper is from 2014! Paper Title The title is short and sweet, at it is called "Generative Adversarial Nets". Everybody who
Read more