Feedbacks on Project Slides

General Comments: A few comments apply to multiple project slides. These comments are summarized below:

1. For a 15 minutes talk, please limit number of the slides no more than 10-15 (ideal would be 10 slides)


2. Ideal font size would be 20 or 24. Font size should be consistent in all slides.


3. In most of the presentations Outline slide is missing. If would be nice if you have after the title slide (1st slide)


4. Overall the presentation can be in following order:
   - (1 slide) title slide
   - (1 slide)an outline slide (it is there in your slides)
   - (1 slide)Motivation and Key challenges slide.
   - (1 slide-optional)Motivation example slide, if there is
   - (1 slide)Related work and their limitations slide(can be 1-2 slides)
   - (1 slide)Contribution slide (Which limitations are you working on?)
   - (1 slide)Problem statement slide
   - (1 slide-optional- can be combined with problem statement slide)Example slide for problem statement, if there is. What is input? What is output?
   - (1 slide)Key Concepts slide, if needed.
   - (1-2 slide)Method, proposed approach slide.
   - (1 slide)Experimental setup
   - (2 slide)Validation slides ( You may not report all the experiment results, but you can report the ones that you think is important.
   - (1 slide)Conclusion and Future works slide
   ----- Total 10-14 slides (except title and outline slides)

• After first slide you need an outline slide.
• slide 2: "Need for efficient storage of structured data" this is the challenge of the motivation.
• After motivation, you need why it is challenging slide.
• slide 3 and 4 seems as basic concepts slide. They can be moved after contribution slide.
• For expriment section you did not report any result. Is it possible to show some results or graphs.
• Slide 14: It will be nice to see a flow diagram. Which dataset did you used for the experiments? The example dataset given in slide 3. If so repeat slide 3 after experimental setup slide for completeness.
• In related work: It may be better to list pros and cons of the related work in English? If so, you may not need slide 6 and 7.
• slide 10 and 11: is it possible to decrease the font size and to have a copy of the RDF shema in these slides.
• for query types slide: You may draw a decision tree-like figure for main queries and subqueries of the main queries?
• If you need, decrase the font size.

• Please check general comments
• Slide 3: If possible you can add some nice figures.
  - Use some keywords instead of sentences
  - I think, RDF one of the solution to Semantic Web.
• After slide 3 you need one more slide listing challenges of Semantic Web or RDF data management.
• After slide 3 again. You can add an example slide showing semantic web example and what RDF representation is etc.
• After problem definition slide add one more slide showing problem definition parameters if possible.
• You can limit related work with 2 slides, If you list some of the limitations in the first related work slide.
• Before our approach , you need a contributions slide.
• It seems that slide 17 should given before slide 15.
• In slide 19, you can add an experimental design figure. I assume you will explain the dataset during the presentations. It would be nice to include a figure or some key features of the dataset.
• In Validation slides figures sizes are too small. If you want you can show the important experiment results insteat of showing them all.
  - Add one or two sentence summarizing each experiment results.
• Future work slide can be added.

• Before problem statement slide, you might give motivations and challenges slide. Please check possible slide order comment in general comments
• Slide 3: Give objective and constraints, if possible.
• Slide 4: You can add some figures in motivation slide.
• Why the problem is hard or challenges slide is missing.
• slide 14: Is it possible to have an animation showing how the algorithm works?
• Before slide 7, you can add one more slide showing input and outputs of the system.
• There are too many approach 2 slides. If possible, you can reduce them by including main idea of the approach

• Please check general comments
• Slide 1: Please include group and class details, class id, student id, etc.
• Motivation slide: You may include some nice pictures
• Slide 5: Font size is too small. Use at least 16 font size.
• Figures are too small. The title of each slide covers too much space. It would be better if you could decrease the space of the title. Or use a different template.
• Slide 15: Font size is too small.
• You can combine slide 6 and 8.
• Slide 7: I do not think you need to give that slide since you are comparing both approaches in slide. It is also not clear in this dataset what the dataset is.
• Related work slide: Your main related work is MDCOP hovewer you can also mention about other spatio-temporal co-occurrence patterns. You can say that they do not have a representation of the dataset.
• slide 11: An animation would be nice to see how the algorithm works step by step

• Please check general comments some slides are missing e.g. outline, contributions, related works and their limitaions, conclusion, experiment setupetc.
• Slide 1: Please include group and class details, class id, student id, etc.
• It would be nice to see some text info or keywords in motivation slide.
• Please try to apply presentation outline given in general comments above, e.g. Challenges can be mentioned before problem statement.
• You can include a flow diagram which shows steps of the system
• If you have, you can add a validation slide or experimental setup slide.

• Please check general comments
• Slide 2-3:
  -Motivation slides 2 and 3 can be combined.
  -In slide 2, some keywords can be pointed out instead of using sentences.
  - If possible, some figures and animations can be added related to presentation and keywords.


• Slide 4: This slide can be given after related work limitations.
• Slide 5: is this slide illustration of motivation? or problem statement.
  - if it is for motivation it should be presented after motivation slide.
  - if it is for problem statement, it can be given after prob. stat. slide to show inputs and output.
• Slide 6 can be presented after the motivation slide.
• Slides 7, 8, 9:
  -2 related work slides would be fine.
  - It possible, you can put a decision tree figure for related work that covers data cube as well as map cube. In the second slide you can list the limitations.
• Slide 11:Contributions should be parallel with the limitations of the related work. For example, which limitation(s) are you covering in your study? Is there any new contributions? etc.
• Slide 12: Can key concepts be explained easily with the lattice structure given in contribution slide? If so, you can use same figure?
  -- You need an approach slide: In this slide you can introduce extention of mapcube model for crash dataset. Then you can mention about experiments.
• Slide 13: Before validadion slide you can list types of queries/questions/parameters that your are trying to visualize can be explained in an experiment setup slide.
• Slide 14-18 For each validation slide, a text can be included to summary the visualization or pattern discovered.

• Please check general comments
• Slide 1: Please include group and class details, class id, student id, etc.
• Outline slide is missing.
• You can include some nice figures in motivation slide
• Please check order of the slides based on general comments given above.
• slide 3 and 5: You can combine these slides
• slide 4 can be an example of problem statement
• slide6:Instead of long sentences, please write some keywords. During the presentation you can explain keywords in detail.
• slide 8, 9, 10, and 13: You can combine these slides using keywords and a related work decision tree. You may na
• Combine contributions slides using keywords instead of sentences
• You need an experimental setup slide, explaining experiment questions, dataset and the process
• Why is it difficult slide should be given after the motivation.
• Conclucions slide: Summarize sentences and use keywords.
• You need a future work slide

Feedbacks on Project Reports (P6)

General Comments: A few comments apply to multiple papers. These comments are summarized below:

1. A rough draft should include following sections (format)
   a. An abstract
   b. Introductions: Problem Statement, Significance of the problem, challenges of the problem
   c. Related Work, Limitations of Related Work and Contributions
   d. Proposed Approach
   e. Validation of listed contribution (experimental, analytical)
   f.Conclusions and Future Work

2. In homework "P6: Rough draft", it is expected to see some execution trace examples of the proposed algorithms/methods and preliminary analytical/ experimental results. It would be better, if you could include your preliminary results and execution trace of your algorithm.
   a. Please look at the Group 8's "P6:rough draft" to see an example of the experimental results, even though format of the draft is not completed.
   b. Please look at the Group 6's "P6:rough draft" to see an execution trace example (see section called materializing the MDCOP Graph). c. Here are three sample project reports from previous csci8715 classes. Please check format and content of these samples while you are revising your project report. Sample Project Report 1 , Sample Project Report 2 , Sample Project Report 3

• See instructor or TA asap to discuss how you would proceed your project based on the comments.
• There is no execution trace and experiment results. See item 2 in general comments.
• Apply draft format given in general comments item 1. Experimental results, conclusions and future work sections are missing.
• To classify or to show the differences of the related work, a decision tree or a table can be given.
• Please justify why you are only comparing path-based and vertically partitioned RDF databases. Why not other methods listed in related work.
• Clearly identify what the challenges of this problem are. If necessary create a new sub section under introduction section.
• Explain in detail, how the task list will be completed.
• Finalize finding benchmark queries. One solution might be using same queries used in the related work.
• List experiment parameters. Which parameters of the methods are you planning to compare/test/evaluate?
• Is it possible to give an algebraic cost model?

• See instructor or TA asap to discuss how you would proceed your project based on the comments.
• Please write the paper in sections. Apply draft format given in general comments item 1.
• There is no execution trace and experiment results. See item 2 in general comments.
• What is the validation methodology that will be used in the paper? Are planning to compare with other algorithms?
• To classify or to show the differences of the related work, a decision tree or a table can be given.
• Clearly identify what the challenges of this problem are. If necessary create a new sub section under introduction section.
• List experiment parameters. Which parameters of the methods are you planning to compare/test/evaluate?

• See instructor or TA asap to discuss how you would proceed your project based on the comments.
• There is no execution trace and experiment results. See item 2 in general comments.
• What are connection and diameter queries? Please explain why you are focusing on connection and diameter queries?. Discuss whether there are any methods in the literatures to answer there queries.
• The paper aims to investigate graph based storage techniques for RDF. However it does not discuss whether there are any graph databases or indexing techniques?
• Graph storage techniques should also be discussed in terms of graph databases in the related work section.
• Why CCAM is chosen is not clear. Is there any other graph based indexing techniques?
• Please complete give preliminary results of the experiments.

• See instructor or TA asap to discuss how you would proceed your project based on the comments.
• Please complete abstract
• To classify or to show the differences of the related work, a decision tree or a table can be given.
• It seems that related work section focuses on specific papers. Are they representing families of related work? If it is possible an example/figure can be given for each class of the related work.
• In comparison section, related works were compared based on several criteria. How would you justify whether the comparison is correct or not since methods are mostly sensitive to the availability and characteristics of the dataset? Is there any benchmark dataset or example to compare related work?

• See instructor or TA asap to discuss how you would proceed your project based on the comments.
• How is this project relevant to the csci8715 class? Please justify.
• Please write the paper in sections. Apply draft format given in general comments item 1.
• Please use consistent font and font size.
• Related work only focuses on one paper. Expand related work by covering subsequence alignment studies and by discussing pros and cons of these methods. Finally you can focus on parallel approaches.
• Related work contains to much technical details. You can give details of the algorithm in the basic concepts or proposed methods section.
• Please complete experiments.
• References should be included in the paper.
• What is range-tree? Is it a spatial indexing tree. If so, It may be better to apply range-tree instead of parallel implementation of the algorithm.

• See instructor or TA asap to discuss how you would proceed your project based on the comments.
• Expand related work by including other spatio-temporal pattern discovery techniques for multiple interest measures.
• Discuss applicability of classical mining methods handling different user constraints for your problem.
• There are no experiment results. See item 2 in general comments.
• Please justify: Why do you need to use TAG representation. Discuss pros and cons. Are there any other methods/techniques to solve your problem?
• Pseudo code is too long? Break pseudo code into functions and explain each function briefly in the execution trace.

• See instructor or TA asap to discuss how you would proceed your project based on the comments.
• Apply draft format given in general comments item 1.
• There is no execution trace and experiment results. See item 2 in general comments.
• Most of the sections are missing. Please include related work, validation methodology.
• Please collect dataset or build a small dataset for the experiments asap
• Define your queries, discuss importance of these queries and give details of the experimental evaluation.

• See instructor or TA asap to discuss how you would proceed your project based on the comments.
• Apply draft format given in general comments item 1. Make bold section titles
• Include a paper title and author names.
• Please complete introduction section.
• To classify or to show the differences of the related work, a decision tree or a table can be given.
• Experiment results seem nice.
• Before experiments please discuss crashcube methods and identify novelties. If necessary, include a key concepts section.

• See instructor or TA asap to discuss how you would proceed your project based on the comments.
• Break introduction section as related work, problem definition, contributions and scope. Figure 1 belongs to related work section?
• Paper directly focuses on time aggregated graph. Is there any other method other than time aggregated graph in the literature to solve your problem?
• To compare related work use same dataset. Currently Figure 2 and Figure 3 have different datasets.
• Include an application domain example reflecting your problem. Discuss it in the introduction in a paragraph.
• Section 2 can be part of related work. Time aggregated graph details can be discussed in the basic concepts section.
• Please explain extended model in detail. If necessary, break the algorithm into functions and explain each function.
• Please complete experiments.
• In the experiments, CCRD and CCRP-NEPTS will be compared. In that case, do you need to explain time expanded graph. Why do not you compare with time expanded graph extension of CCRP.

• See instructor or TA asap to discuss how you would proceed your project based on the comments.
• Paper includes one interview.
• How will you evaluate results of the interviews to come up a conclusion? Can  you explain final step in section 5 in detail.
• Can you check literature how smoke detectors or fire sprinkler affected insurance value of a house? It might be relevant.

Feedback on Project Proposals (P4,P5)

General Comments : The proposal (P5) is a union of the answers to questions P1, P2, P3 and P4. This you should include the list of readings, summary of readings from last homework in the proposal instead of creating new ones.

1. Two sample proposals are given in the class web page. A proposal generally contains following sections.
1. An abstract
2. Research objectives or Project description or problem statement
3. Related work or Literature survey and novelty of the project
4. Research methodology and research tasks
5. Conclusions and future work
6. References

G1 -----------------------------------------------------

• Add an abstract for the proposal.
• Overview or introduction section should be more general. Currently it focuses on two related work. It should focus on an introduction for “What is RDF”, “Why is it important”., “Why do we want to store them”, “What are the challenges to store them, e.g, query answer time, number of joins, etc. ”. Then you can list objectives. Currently this section is kind of related work. For introduction section you can use importance and relevance section.
• Related work: Currently it seems that you have two classification of the related work: 1) Vertical partitioning vs path-based storage and 2) Normalized triple store vs dynamic table schema. However the connection between these classifications is not explained. The best way to solve this problem is to pick one of the related work classifications and to use the second one as a sub-class of the first. It would be better if you could give a decision tree figure for the classification of the related work.
• Use figures to explain terms used in the Related work section. You might even put a Basic concepts section to explain basic terms.
• Add a research methodology section?
• Task list: For each task: Can you explain it in a couple of sentences? Why is the task important or why did you select this task? And how do you want to solve it.
• There must be several benchmarks to test performances of the classical databases. Is it possible to use one of them to test performance of the RDF storage methods.? Why or why not.
• First page, second paragraph: There is a typo , “nodes are co Both approaches”

G2 -----------------------------------------------------

• P4 and P5 have not been submitted yet.

G3 -----------------------------------------------------

• Proposal should have a title.
• P5. Formal proposal : Includes questions or thesis, key resources, and proposed structure revised with feedback from peers and the instructor. The proposal should list a methodology to validate the claims made in the contributions section. Choose a methodology similar to the ones used in the papers in the reading list. For example, experiments, analytical methods, case study, detailed illustrative examples, prototyping and demonstration of new capability, etc. There should be a clear plan to list the steps within each methodology. For example an experimental methodology should include a description of the experiment design listing the candidates to be compared, metrics of performance, values of fixed parameters, valuesets for variable parameters, benchmark datasets and computations, key assumptions, etc.

G4 -----------------------------------------------------

• As noted in P3, it would be better, if the scope of the survey is narrowed.
• A classification of the related work should be given. Classification can be done based on the assumptions of the related work or the problems they are trying to solve.
• Based on the classification of the related work, validation methodologies should be explained. For example, experiments, analytical methods, case study, detailed illustrative examples, prototyping and demonstration of new capability, etc. There should be a clear plan to list the steps within each methodology. For example an experimental methodology should include a description of the experiment design listing the candidates to be compared, metrics of performance, values of fixed parameters, valuesets for variable parameters, benchmark datasets and computations, key assumptions, etc.

G5 -----------------------------------------------------

• A formal definition of the problem can be given in terms of DNA sequence alignment problem. What is given, what do you want to find, what is objective, what is constraint? The problem statement given in “previous methods” is the problem statement of your problem or previous methods?
• Introduction section should explain: 1) Why is the problem important 2) Why is it challenging questions. Application domains of the problem can be listed. Examples of the problem can be given. You can use some text from P3 part of your text  http://groups.google.com/group/spatial-z/web/project-page
• Currently introduction section is kind of related work section. It discusses the LAGAN software. Introduction section should be rewritten.
• Related work should not be limited to LAGAN software. A general discussion of the solving problem of DNA sequence alignment discussion should be given. Limitations of the related work should be explained.
• If possible do not give pseudo code of the algorithm and explain it in words (may be in a paragraph). If you need you can refer to the pseudo code of the algorithm. Some of the notations of the algorithm is not explained.
• What are the limitations or bottle necks of the serial algorithm? Which function or which step you want to parallelize?
• It might also given why current spatial indexing structures (e.g. R-tree, quad tree, etc) are not suitable for the problem?
• What kind of parallelization method , you will be using (e.g. OpenMp, MPI, etc) and why?
• Discussion of the methodology can be given. What are the experiments you are planning to conduct? Which parameters are important for your problem?

G6 -----------------------------------------------------

• It would be better if you could classify the related work based on their assumptions or properties. You can consider giving a decision tree or a table that classifies related work.
• Contribution claims and experimental validations can be narrowed if possible.
• The details of the methodology for each contribution claim should be given.
• The question of “why the problem is important” is not addressed clearly.

G7 -----------------------------------------------------

• It would be better, if you could give a classification of the related work.

G8 -----------------------------------------------------

• It would be better, if you could give a classification of the related work and their limitations

G9 -----------------------------------------------------

• Details of the validation methodology and experiment should be discussed.
• Algorithm design decisions and/or proposed new solutions should be discussed.
• In formal proposal P5 you can combine all your projects (P1-P4) you wrote.

G10 -----------------------------------------------------

• In formal proposal P5 you can combine all your projects (P1-P4) you wrote.
• Give a classification of the related work
• Research and validation methodology discussion should be given.

Feedback on Paper Analysis Slides
General Comments : A few comments apply to multiple papers. These comments are summarized below:

1. A cover slide should include a citation (e.g. title, author, publication forum, year, etc.) to the paper you are presenting. Cover slide should also include information about the team, including the names of members, group's webpage url, etc. Motivation slides should either list application domains or long-standing open problems in the field of spatial databases.
2. Use phrases instead of complete sentences. Each phrase should be 6 to 8 words. You may omit articles and reword phrases to reduce the number of words. Remember the slides are providing highlights and you would be able to fill in the details in the oral presentation.
3. Limit the amount of text on each slide to 6 to 8 phrases. Too much text forces the audience to read the text and prevents them from listening to you. Consider covering parts of the slides with too much text during presentation.
4. Put a slide on "outline" listing the major groups of slides. Know the number of slides in each group. Time your presentation to know the amount of time you spend in presenting the slides in each group. This will allow you to complete the presentation on time.
5. Problem statement slide should identify four things:
• inputs (what information is given),
• outputs (what solution is to be found ),
• objectives (how does one judge the quality of solution), and
• constraints (which assumptions can be made on inputs and environment).

Problem definition should be stated in a way that allows solutions other than what authors have proposed. This is fundamental to understand and critique any paper without being biased by the assumptions/biases of the authors. Many research papers will have a section on problem statement identifying these components. Others will list assumptions in scope, experiment design, proofs of theorems and lemmas, proposed algorithms/methods etc. If this information is missing, you should try to come up with best guesses for these components.

6. Key concepts should be defined and explained. Consider using examples and figures in explaining the key concepts. Use examples from paper. In addition you may add new examples. You may refer to figures in the papers by proving figure numbers and page numbers. Figures should be simple with few (e.g. half a dozen) components. Consider using 3 to 5 slides on key concepts since audience should understand these well to appreciate rest of your analysis.
7. A validation methodology allows researchers to back up their claims in an objective manner. It allows readers to reproduce the results and conclusions of the authors. This is a fundamental requirement for science and scientific methods. Common methodologies used in computer science research include the following:
• Software prototypes and implementations for use by a user community. Papers often quote popularity of the system and testimonials from users. Example include first implementations of Unix, Ingress, mosaic, terraserver (paper 21), Geominer (paper 3) etc.
• Experimental evaluation of design decisions using a prototype of a simulation model (e.g. queuing model, algebraic cost models). It is useful to provide some details of the experiment design, e.g. benchmark data/queries, fixed/variable parameters, candidates evaluated and simulation model.
• Theoretical proofs of correctness, computational complexity. Nature of the proof (e.g. contruction, contradiction, reduction...) may be discussed.
• Examples, detailed examples, case studies of real usages are used for concepts where more formal validation may be difficult or too early to do. Data modeling papers often use this. Surveys of experts and statistical analysis
8. Assumptions, Rewrite Today: It is good to add these two slide to distinguish between the conclusions of authors and your conclusions.

Comments on individual groups
G1 -----------------------------------------------------

• The slides are summary of the book, however it did not address critical reading questions. Basically 6 items in critical reading is missing. They are listed below. Please also check general comments.
• Problem definition slide is missing. See item 5 in general comments.
• Motivation (Why is it important?) slide is missing. Application domains which will benefit from a solution to this problem.
• Why is it hard? slide is missing. Try to identify novelties of the proposed algorithms listed with regard to the related work
• List of contributions slide is missing.
• Novelty of contributions slide is missing.
• List of assumptions slide is missing
• List of revisions slide is missing. See item 8 in general comments.

G2 -----------------------------------------------------

• Break problem statement slides into bullets. Please see item 5 in general comments.
• Make a separate slide for why is it difficult part. Second part of problem statement slide.
• Label axis of graphs in slide “bifurcations”
• Slides generally include long sentences. Please break them into bullets. Please see item 2 in general comments.
• Add a slide about key contributions of the book and to the field of economics. Identify contributions of interest to people in CSCI8715.
• Add a slide on rewrite today. See item 8 in general comments. How will you revise it if you rewriting it today.

G3 -----------------------------------------------------

• Add slides on why is the problem important and why is it hard?
• Add a slide on in what sense proposed approach is novel related to related work?
• List experiment results and show why proposed method is better.

G4 -----------------------------------------------------

• Comments provided by email earlier since they are presenting first.

G5 -----------------------------------------------------

• Slides are summary of the paper but it did not address critical reading question. Please see general comments and modify slides based on 6 items of critical reading.
• Add and/or make separate slides about problem statement, why is the problem is important, why is it hard, assumptions, contributions, and rewrite today.
• Instead given pseudo of the algorithm, list key ideas of the algorithm.
• Add experiment design and experiment goals information before the experiments slides.

G6 -----------------------------------------------------

• Add slides on motivation,  why is the problem hard, 
• In contributions slide, try not to use long sentences.
• In illustration slides, there are too much text. If necessary break slide into 2-3 slides.
• Add slide that is showing experimental results, comparison of the algorithm.
• Add on list of revisions if you were to revise the paper today.

G7 -----------------------------------------------------

• Revise problem statement slide as explained in general comments.
• Break long sentences into bullets.
• In key concepts slide (3^rd slide): Explain or illustrate how multimodal models are supported
• Last key concepts slide seems to sparse add figure or additional text to provide details.
• Add separate slides on why is the problem important, why is it hard (e.g. heterogeneities across multimodal networks), and contributions (e.g how did they address heterogeneity across multiple modes of transportation.
• Add slides on key assumptions made by authors, list of revisions if you were to revise the paper today.

G8 -----------------------------------------------------

• Add information to the first slide. See general comments.
• Add slides on problem definition,  why is the problem hard, and revisions you were rewrite the paper today.

G9 -----------------------------------------------------

• Revise problem statement slide as explained in general comments.
• Break long sentences into bullets
• List limitations of related work
• Instead of pseudo code of the algorithms, provide key ideas of the algorithms
• Add a slide on experimental setup and experiment goals before experimental results.
• Add slides on assumptions and, list of revisions if you were to revise the paper today . 
• Add a slide on why is the problem hard.

G10 -----------------------------------------------------

• Add slides on problem definition, why is this problem hard.
• Break long sentences into bullets (e.g slides introduction, threat, solution proposed, working of GSAM model, assumptions, examples)
• Add slides on revision if you were to rewrite it today.