Final Project: Evaluating a Research Paper

The link of my evaluating paper:

http://www.cs.berkeley.edu/~whitney/research/Whitney_Isailovic_Patel_Kubi_Frontiers007.pdf

1. Describe the audience for whom the paper is written. (Provide detail and do not assume that the audience is homogeneous).

a. What concern of the audience is addressed in the paper?

b. Is this concern primarily practical or theoretical?

Audience: Researchers those who had been or will be studying “Quantum Computing” technologies, especially those try to use CAD (computer-aided design) in designing the quantum circuits.

a. Automate most of the tasks involved in generating a physical layout and its associated control logic from a high-level quantum circuit specification.

b. This concern is practical.

2. How is the paper structured? How many headings does the paper use and how are these formatted in the paper? Does the paper follow the APA headings style (http://owl.english.purdue.edu/workshops/hypertext/apa/parts/headings.html)?

The structure of this paper:

Section 1: Introduction

Section 2, 3: Background Knowledge

Section 4, 5: Methodologies – Common method used to control the experimental environment

Section 6, 7, 8: Methodologies in detail – Techniques used in this research

Section 9: Results

Section 10: Conclusions

The paper uses two levels of headings, and the format is:

8. DATAFLOW-BASED LAYOUTS (Level 1)

8.1 Dataflow Graph Analysis (Level 2)

8.2 Placement and Routing (Level 2)

8.3 Annotated Scheduling (Level 2)

9. RESULTS (Level 1)

9.1 Benchmarks (Level 2)

9.2 Evaluation (Level 2)

This paper does not follow the APA heading style. It follows the ACM style.

3. Does the paper contain an abstract? What kind of abstract is it? (See AWGS p. 282).

Yes. The abstract in this paper is results-driven Abstract.

4. What is the paper’ argument structure?

a. What is the paper’s main claim?

b. What reasons does the author provide to support the claim?

c. What type of evidence does the author provide?

Provide some examples of evidence.

Is the evidence quantitative or qualitative?

d. Does the author acknowledge counter-arguments? If so, how does the author

deal with the counter-arguments?

e. Does the paper contain any warrants?

a. Classical electronic circuits can be designed using the CAD (computer-aided design). So can quantum circuits.

b. Provide the quantum CAD flow and compare it with the classical CAD flow.

c. The authors conduct the experiments and get the data. Those data contains the latency and area of the designed circuits. All the data can be provided as evidence. And also, the evidence is quantitative.

d. No.

e. No. It’s about hard science and no such warrants exist.

5. Which moves does the author use in the introduction? (See Table 6 on AWGS p. 244). Identify the moves and describe them. What words or phrases mark the beginning of each move?

In the introduction, the author uses move 1, 2 and 3. The phrases “Quantum computing” in the first sentence marks the beginning of move 1, and the word “while” in the last sentence of paragraph 2 marks move 2. The words “Our goal” in the first sentence of paragraph 3 marks move 3.

6. Does the paper contain any definitions? If so, is it a gloss, a sentence definition, or an extended definition? (See AWGS p. 50-51).

No definition.

7. What is the author’s method? Using Table 22 on AWGS p. 227, give the methods section a score. Does the author justify her/his methodology/

The methods the authors mentioned in the paper are very technical, and they provided 3 kinds of methods to control the experiments. If I follow Table 22 on AWGS p. 227, I will score the method section -6. It seems that the authors didn’t justify their methods.

8. In the paper’s results and discussion sections, what examples of hedging (i.e., qualifications) do you see? (See AWGS p. 125). List at least five examples.

In fact, in this paper, not many examples of hedging can be seen. And, I try to find 5 examples which may be not good. 1^st: The exhaustive search over grids yields the best latency for all benchmarks. 2^nd: Dataflow-based place and route techniques in general produce the lowest latency circuits. 3^rd: We find that greedy heuristic tends to find the best design area-wise. 4^th: Our greedy algorithm produces good result for very simple circuits. 5^th: We found that our dataflow approach generally offers the best latency, often at the cost of area.

9. Which moves do you see in the Discussion Section? (See AWGS p. 270.)

In fact, this paper did not have a “Discussion Section”, but in the “Conclusion” section, the authors mentioned all of Move 1, 2, 3 in that section.

10. Does the author evaluate any previous research? If so, what evaluative language is used? (See AWGS, p. 188).

The author evaluate two previous research paper, discuss the method they use in designing and optimizing the circuit, and then the sentence “While this approach is quite informative in a new field, it quickly becomes intractable as the size of the circuit grows.

11. How does the author refer to the end-of-text citations? How current are most of the citations? If there are older citations, what do these refer to?

The author uses [number] format to refer to the end-of-text citation (IEEE or ACM format maybe). Most of the references cited in this paper were published in 10-15 years. The reason of this is that the “Quantum Computing” technologies were not researched before 15 years. Only one research paper cited in the article was published more than 20 years ago. This paper was cited because the theory described in it is the fundamental theory used in the article.

12. Create a summary (125-200 words) of the paper.

In this paper, the authors try to point out the most important tasks in designing and optimizing a quantum circuit and try to find one or more possible solutions by using computer-aided design (CAD). After describing the background knowledge needed to conduct the research and the common method used to control the experimental environment, three group of researching object were introduced in the paper, such as “grid-based layouts,” “greedy place and route,” and “dataflow-based layouts.” After the results of these experiments emerged, comparison of these original data was done. According to the comparison of the result, no globally optimal method was found, but each method was found to be optimal in some special circuit designing condition. Also, when doing this experiment, some possible future study was found.