Learner and Observer, be happy to share useful things.

How to use artificial intelligence services (such as ChatGPT) for efficient academic research? (2024)

Before the use of large-scale language models in artificial intelligence, Google Scholar was widely used for academic research. However, with the introduction of AI systems like ChatGPT, various AI-powered research assistant services have emerged that utilize academic data and standards to train large language models, providing more specialized support for researchers compared to Google Scholar. Examples of such services include Elicit and Semanticscholar, which seem to better cater to the needs of researchers. In this article, I will outline the currently available AI services and how they can be used for academic research.

1. Available AI Services for Big Data Models#

1.0 Microsoft Copilot. If you are in the US or similar regions, you can use the GPT-4+bing real-time exploration information model for free.

To use Microsoft Copilot, there are two main methods: first, the Copilot plugin on the upper right side of the Edge browser on your computer; second, installing the Microsoft Copilot app on your mobile device, which is currently available for Android and iOS. Additionally, you can apply for Github Copilot and install the GitHub Copilot Chat plugin in VSCode to temporarily use the GPT-3.5 model for free.

1.1 ChatGPT by OpenAI, the free version allows the use of the ChatGPT3.5 model.

1.2 Use Google Gemini service. Gemini.

1.2 Use poe.com. It provides various models such as ChatGPT and Bard for free use. You can temporarily use the ChatGPT-3.5 model (more convenient than logging in to the official website).

1.3 Coze.com developed by ByteDance.

1.4 Use article-related artificial intelligence (AI) services.

1.4.0 https://www.scinapse.io; (recommended) Claims to be more powerful than Google Scholar (https://scholar.google.com) and philpapers (https://philpapers.org).

1.4.1 https://old.elicit.org; (recommended)

1.4.2 https://www.semanticscholar.org; (recommended)

1.4.3 https://openknowledgemaps.org; (recommended)

1.4.4 https://researchrabbitapp.com;

1.4.5 https://oa.mg; (recommended)

1.4.6 https://www.researchgate.net; (possible)

1.4.7 https://www.researcher-app.com; (registration required, not tried)

1.5 Large-scale AI services for reading PDFs and answering questions.

1.5.1 https://www.scinapse.io; (recommended)

1.5.2 https://www.readcube.com;

1.5.3 https://sider.ai/chatpdf or https://www.chatpdf.com/; (recommended)

1.5.4 https://www.readcube.com; (recommended)

2. How to Use AI Services for Big Data Models#

2.1 First, identify the problem and its general outline or structure. This can be done using Google (https://www.google.com/ncr) or the Bard service (https://gemini.google.com).

2.2 Next, read core materials related to the problem. Core materials can be discovered through citation analysis or by referring to texts frequently studied by prominent philosophers.

2.3 Third, familiarize yourself with the content of the core materials.

2.3.1 You can have AI read the PDF to provide an outline or general argument structure of the article.

2.3.2 Read the article yourself and delve into the argumentation and potential issues. (AI may make mistakes)

2.3.4 Have AI read the PDF and identify potential issues with the article.

2.4 Read related articles.

2.4.1 Use AI services for finding related articles.

2.4.2 Use AI services for reading and summarizing related articles, and after manual review, provide accurate conclusions or viewpoints.

3. Conducting Research, Writing, and Submitting Manuscripts Based on the Problem#

3.1 Start with the core viewpoint and expand in both directions.

3.2 The literature review can refer to the references and related literature suggested by AI (must be manually reviewed).

3.3 The core argument can refer to the argumentation provided by AI.

3.4 The argumentation provided by AI can be used as a reference but should not be the sole basis of the argument.

3.5 Complete the writing and revision of the paper, submit it to the journal editor, make revisions based on the editor's requirements, and ultimately get it published.

Note: This article was written on the evening of December 31, 2023, and published on January 1, 2024.

For advanced usage of ChatGPT, refer to: Source1, Source2.

Ownership of this post data is guaranteed by blockchain and smart contracts to the creator alone.