Literature retrieval practice

Literature research practice

Do a systematic and comprehensive survey based on the documents given by the teacher.

Comprehensive and systematic search "memristor"

1. Wikipedia retrieval stage: In order to understand the concept, we first search for "memristors" in Wikipedia to get a rough understanding of the development of memristors. The following information is obtained: Leon Chua (1971, IEEE) first proposed this concept, and then he and his collaborators proposed the "memristor system" (1976, IEEE); started the research on memristor in the 21st century and promoted memristor The climax of the research on the device is the article published by the researchers of HP Labs (2008, Nature). In addition, we can also pay attention to the articles mentioned in the entry (early), which generally involve the nodes of memristors development, so that we don’t need to pay attention to each article when we look up early documents later (so because Finding early documents relied on the EV platform, because the WOS core set began to be included in 1991, but the EV platform does not have a citation sorting function, so that we cannot determine hot articles by citation sorting in a limited early time period) . We learned that in the early articles, in addition to Leon Chua's article, there are some articles. Through References under the Wikipedia entry, we can link to the above mentioned documents. Of course, this cannot be used as our main way to obtain literature.

2. Database retrieval stage:

   When we search for documents, we pay attention to two stages of documents in terms of time. One is the earliest document, which is generally the pioneering work of this concept; the other is the most recent document, which can reflect the frontier dynamics of the development of the concept.

• Early literature: By reading the earliest literature in a field, we can learn how to associate, how to propose new concepts, and how to innovate; in addition, we can form some ideas of our own by looking at these literatures first, so that we can compare our own ideas when looking at the follow-up literature. Compare ideas and learn how to find breakthrough points.

  • For memristor, we learned that the concept was first proposed by Leon Chua (1971, IEEE) in the search stage of Wikipedia, so we searched for memristor in IEEE Xplore. The time limit was 1971-1980. Sorted by Most Cited (Paper), we found the first two articles. The citations of both are quite high, so both articles have been preserved.

  • However, there is a characteristic of the development of memristor. Although Leon Chua proposed the mathematical model of memristor as early as 1971, the climax of the research came from the article published on nature by the researchers of HP Lab in 2008, so we still Pay attention to the following article. Since it is in nature and in 2008, we will download it from WOS in the scope of WOS.

• Recent literature: Search the literature of the last 1 to 2 years, sort by the number of citations, and select the highest ones. You can understand the most cutting-edge research focus of this concept. However, the literature found by this method may only focus on a certain research. Point (for example, focus on using memristor to build AI hardware platform), so the research points of other concepts (such as memristor's material characteristics) that appear first in the ranking of citations. In addition, the WOS screening system can filter the types of documents, and we can filter the types of reviews. Since memristor is a fast-developing concept, the timeliness of review articles is very important. We limit the time to the last two years. Sort by the number of citations, and select the top few reviews. Similarly, in order to prevent the review from being limited to only one research hotspot (such as neuromorphic computing), you should not only focus on ranking, but also different concepts under memristor. Of course, it would be better if there is an overview of all aspects of memristor.

  • On the WOS platform, create a citation report for the search results under memristor. The limited time range is 2019-2021, and the document type is not limited. Sorted according to the frequency of citations, the following results are obtained:
    

    Analyzing the above results, we can notice that some research points focus on the realization of artificial intelligence platforms, and some focus on preparation and implementation (the issues related to research points will be discussed at the end).

    The time we searched above is from 2019 to 2021, and the most cited articles were published in 2019. Thanks to the advantage of earlier time, this will cover 2020 closer to us, so we will limit the time range to 2020- In 2021, the types of documents are not limited, sorted according to the frequency of citations, and the following results are obtained:

    The hot articles in the last two years need to be paid attention to. Download the above 7 articles (the number is not finalized, there may be other less popular but promising research sites that have not been paid attention to, and the retrieval of specific aspects will be discussed at the end).

  • We further search for review articles, use the filtering function of WOS to limit the document type to review, and the time limit in the citation report to 2019-2021, sorted by the frequency of citations:
    

    The three reviews focus on: application to AI, material properties (perovskite), and physical application (Wayne Bridge Oscillator).

    Then limit the scope to 2020-2021, sorted by the frequency of citations:
    

    Note that the first article is not directly related to memristor, but discusses neuromorphic spin electronic devices. Memristor is just one of the keywords, so the correlation is not too strong, so we focused on the next three articles: Used for Edge computing (AI field), physical characteristics, based on specific materials (graphene).

3. In-depth retrieval stage: Our previous retrieval only selected the first 3-4 articles after being sorted by the frequency of citations, which can only reflect the most popular articles under the grand concept of memristor. In fact, there are many research directions under this concept. For example, it is applied to AI, material realization, etc. Our future research is also in a more refined sub-field. Therefore, to determine other concepts related to memristor, we will use the Engineering Village platform at this time, due to the expansion Inspec Analytics can analyze the controlled vocabulary related to memristor, as shown in the following figure:
   we can perform further searches in the specific direction. Since only the initial stage of retrieval is involved, and many basic knowledge have not yet been learned, no further retrieval is required.