مطالعه بهتر و سریع‌تر مقالات با کمک هوش مصنوعی

دسته‌بندی: متفرقه

آرش صمدی

ایجاد: شنبه 29 دی 1403 13:30

بروزرسانی: جمعه 19 بهمن 1403 13:13

امروزه هوش مصنوعی از کلیدی ترین کلماتیست که در هر حوزه ای که نگاه کنیم موارد کاربرد آن به چشم میخورد. هرچند که این پدیده نو ظهور نگرانی هایی را در میان عموم برانگیخته اما به طور قطع میتوان گفت که تلفیق هوش مصنوعی در امور روزمره همچون قطاریست که نباید از آن جا ماند. در این مقاله به نمونه ای کوچک از کاربردهای هوش مصنوعی Copilot از کمپانی مایکروسافت در بررسی و مطالعه هرچه سریع تر مقالات خواهیم پرداخت.

مطلب پیش روی عزیزان قدری با محتویات دیگر این وبسایت متفاوت است. مقالات این سایت اصولا حاصل مطالعات و گرداوری مطالب از منابع مختلف بوده، اما این مقاله حاصل تجربه شخصی بنده در استفاده از هوش مصنوعی در جهت تسریع بررسی مقالات است. مرورگر Edge از کمپانی ماکروسافت (که به طور پیشفرض در ویندوزهای 10 و 11 نصب شده است.) قابلیت های مختلفی داشته و علاوه بر امکان جستجوی وب، امکان خواندن فایل‌های PDF را نیز برای کاربر فراهم می آورد. علاوه بر این، چندیست که کمپانی ماکروسافت، Copilot، یکی از پیشرفته‌ترین ابزارهای مبتنی بر مدل GPT را به صورت رایگان در اختیار کاربران خود قرار داده است. در سال‌های اخیر کمپانی ماکروسافت در حال اضافه کردن کوپایلت به نرم‌افزارهای متعددی نظیر مرورگر Edge، نرم‌افزارهای آفیس و سیستم عامل ویندوز بوده است. پشتیبانی از فایل‌های PDF و تلفیق هوش مصنوعی Copilot با Microsoft Edge امکانی فوق‌العاده برای بررسی هرچه سریع‌تر و دقیق‌تر مقالات، کتب، استانداردها و هر مستندی را فراهم می‌آورد. 

در شکل 1 از مرورگر Edge استفاده شده تا فایل PDF یک مقاله مطالعه شود. در این تصویر و با کلیک بر روی آیکون کوپایلت (مشخص شده با حرف A در تصویر) پنل مربوطه در سمت راست پنجره مرورگر باز میشود. توجه داشته باشید که از این جهت که سرعت تغییرات در دنیای هوش مصنوعی بسیار زیاد است، ممکن است به زودی و در نسخه های بعدی هر یک از نرم افزارهای کوپایلت یا Edge تغییرات ظاهری بزرگی اتفاق بی افتد. 

شکل 1- تصویری از مرورگر Microsoft Edge، یک فایل PDF باز شده و پنل کوپایلت نیز در سمت راست تصویر قابل رویت است.

در این مثال، کوپایلت به صفحه ای که در تب مرورگر باز شده دسترسی داشته و کاربر میتواند از آن در رابطه با محتویات این تب سوالات خود را مطرح نماید. محتویات تب هم میتواند یک صفحه از وب بوده و هم یک کتاب یا مقاله در فرمت PDF باشد. در شکل 1 در تب فعلی فایل PDF یک مقاله باز شده است. این بدان معناست که کوپایلت علاوه بر این که به فایل این مقاله دسترسی دارد، کاربر میتواند در رابطه با آن سوالاتی مطرح نماید. برای تسهیل کاربری همانطور که در شکل 1 مشخص شده، افزونه کوپایلت در موارد مشخص شده با حرف B (Create a summary) و C (Expand on this topic) دو فرمان از پیش نوشته شده به ترتیب برای ایجاد خلاصه ای از مقاله (B) و توضیح بیشتر در رابطه با موضوع مقاله (C) را دارد. فیلد مشخص شده با E نیز این امکان را فراهم می آورد تا کاربر سوالات و دستورات خود را به صورت کتبی به کوپایلت ارائه نماید. 

مقاله موجود در شکل 1 با عنوان "Adaptive coordination of overcurrent relays in active distribution networks based on independent change of relays’ setting groups" روشی تطبیقی برای حفاظت شبکه های توزیع فعال با استفاده از قابلیت گروه های تنظیمی ارائه میکند. بنده از نویسندگان این مقاله بوده و از این جهت آشنایی کاملی با مطالب آن دارم، اما فرض را بر آن بگذاریم که هیچ چیزی از آن نمیدانیم. هدف آن است تا با استفاده از هوش مصنوعی کوپایلت راجع به آن اطلاعاتی کسب کنیم. از طریق فیلد مشخص شده با حرف E در شکل 1 دستور "Please provide a comprehensive explanation of the methodology and results of this paper" به کوپایلت داده میشود. هدف از این دستور آن است که روش استفاده شده در این مقاله تشریح گردد. پاسخ کوپایلت:

The paper titled "Adaptive coordination of overcurrent relays in active distribution networks based on independent change of relays’ setting groups" presents a novel approach to address the challenges of protecting modern distribution networks with high penetration of distributed generations (DGs) and the capability of network configuration changes.

Methodology

The authors propose a setting group (SG)-based adaptive protection scheme. The methodology involves the following steps:

  1. Network State Identification: The network states are identified based on the presence of DGs and the configuration of the network.

  2. Setting Group Selection: For each identified network state, a proper SG is selected for each relay. This selection is done using integer linear programming (ILP) and particle swarm optimization (PSO).

  3. Optimization: The objective is to minimize the total operating time of overcurrent relays (OCRs) while maintaining coordination constraints. The ILP is used to specify the proper SG for each relay, and PSO is employed to optimize the relay settings within each SG.

  4. Simulation and Validation: The proposed method is applied to a sample distribution network to demonstrate its efficiency. The network is simulated under different fault conditions and configurations to validate the effectiveness of the adaptive protection scheme.

Results

The results of the study show that the proposed SG-based adaptive protection scheme significantly improves the coordination of OCRs in active distribution networks. The key findings include:

  1. Improved Coordination: The adaptive scheme ensures proper coordination of OCRs under various network states, including different DG penetrations and network configurations.

  2. Reduced Operating Time: The total operating time of OCRs is minimized, leading to faster fault clearance and improved protection performance.

  3. Flexibility: The method is flexible and can adapt to different network conditions, making it suitable for modern distribution networks with high DG penetration and reconfigurable capabilities.

  4. Validation: The simulation results demonstrate the effectiveness of the proposed method in maintaining protection coordination and minimizing relay operating times under various fault conditions.

Overall, the paper presents a comprehensive and effective approach to adaptive coordination of OCRs in active distribution networks, addressing the challenges posed by DGs and network reconfigurations.

همانطور که دیده می‌شود، کوپایلت در بخش ابتدایی از پاسخ خود از توضیحات مربوط به روش ارائه شده در مقاله استفاده می‌کند تا ابتدا توضیحاتی کلی داده و سپس از بخش نتایج مقاله برای بررسی مزایای آن استفاده کرده است. اما لازم نیست پرسش را به این نقطه ختم کنیم. حفاظت تطبیقی در شبکه های توزیع، بحثی ریشه دار است، در این مرحله می پرسیم که مهم ترین تفاوت های کار ارائه شده در این مقاله با پژوهش های پیشین چیست؟ برای این کار از دستور "what are the key differences of this paper with the previous research" استفاده میشود. پاسخ کوپایلت به این دستور:

The paper "Adaptive coordination of overcurrent relays in active distribution networks based on independent change of relays’ setting groups" introduces several key differences compared to previous research:

  1. Independent Setting Group (SG) Changes: Unlike previous methods where all relays activate the same SG for a given network state, this paper proposes that each relay independently selects its SG based on the network state. This allows for more precise and flexible coordination.

  2. Optimization Techniques: The paper employs a hybrid approach using Integer Linear Programming (ILP) and Particle Swarm Optimization (PSO) to determine the optimal SG for each relay. Previous methods often relied on simpler optimization techniques or did not combine multiple optimization methods.

  3. Reduced Number of SGs: The proposed method requires fewer SGs to achieve effective coordination. Previous approaches often needed a higher number of SGs, which could be impractical given the limitations of commercial OCRs.

  4. Improved Coordination and Reduced Operating Time: The new method significantly reduces the total operating time of OCRs while maintaining coordination constraints. This is achieved through the independent SG selection and the hybrid optimization approach.

  5. Flexibility and Adaptability: The proposed method is more adaptable to various network conditions, including different DG penetrations and network configurations. Previous methods were often limited in their ability to handle such dynamic changes.

These differences highlight the advancements made by the authors in addressing the challenges of protecting modern distribution networks with high DG penetration and reconfigurable capabilities.

شاید برایتان جالب باشد که این 5 مورد که به زیبایی توسط کوپایلت دسته بندی شده و توضیح داده شده اند، مهم ترین و کلیدی ترین تفاوت های این مقاله با پژوهش های پیشین هستند! هرچند که چند تفاوت جزئی تر نیز در این کار نسبت به کارهای قبلی وجود دارد که درجه اهمیت کمتری دارند. برای رسیدن به این نتایج که کوپایلت در چند ثانیه رسیده، لازم بود تا مخاطب کل مقاله را با دقت مطالعه نموده و با اصل بحث نیز آشنایی مناسبی داشته باشد. 

اما کار به این جا ختم نمیشود. یکی از اصلی ترین چالش ها برای پژوهشگران تازه کار این است که در جستجوهای خود ممکن است مقالاتی را پیدا کنند که لزوما از ژورنالها و مجلات معتبری نیستند. برای این که بدانیم آیا این مقاله در یک مجله معتبر چاپ شده یا خیر از دستور "is this paper from a respected and trusted journal? what are its metrics" استفاده شده است. پاسخ کوپایلت:

Yes, the paper "Adaptive coordination of overcurrent relays in active distribution networks based on independent change of relays’ setting groups" is published in the International Journal of Electrical Power and Energy Systems, which is a reputable and well-regarded journal in the field of electrical engineering and power systems.

 

Key Metrics:
Impact Factor: 6.53

h-index: 165

SCImago Journal Rank (SJR): 1.711

Quartile: Q1 in both Electrical and Electronic Engineering, and Energy Engineering and Power Technology

These metrics indicate the journal's high influence and reputation in the academic community.

میبینیم که بدون نیاز به جستجوی بیشتر کوپایلت اعتبار مقاله و Q1 بودن مجله را تایید نموده و حتی پارامترهای ارزیابی اعتبار مجله را عنوان نموده است. اگر برایمان اصطلاحاتی نظیر impact factor، h-index و امثالهم نیز تازگی دارد و در رابطه با آن اطلاعات چندانی نداریم، با استفاده از دستور "what does each metric mean" در ادامه مکالمه با کوپایلت میتوانیم نسبت به پرسش در رابطه با معنای هر یک اقدام نماییم. پاسخ کوپایلت:

 

Sure, here’s a breakdown of what each metric means:

Impact Factor

The Impact Factor measures the average number of citations received per paper published in the journal during the preceding two years. It is often used as a proxy for the relative importance of a journal within its field. A higher impact factor indicates a more influential journal.

h-index

The h-index is an author-level metric that attempts to measure both the productivity and citation impact of the publications of a scientist or scholar. For journals, it reflects the number of articles (h) that have received at least h citations. A higher h-index indicates a more impactful journal.

SCImago Journal Rank (SJR)

The SCImago Journal Rank (SJR) is a measure of the scientific influence of scholarly journals that accounts for both the number of citations received by a journal's articles and the prestige of the journals where such citations come from. It is similar to the Impact Factor but also considers the quality of the citing journals.

Quartile

Journals are often ranked into quartiles (Q1, Q2, Q3, Q4) based on their impact factor or SJR within their specific field. Q1 represents the top 25% of journals, Q2 the next 25%, and so on. Being in Q1 indicates that the journal is among the most prestigious in its field.

These metrics help researchers and academics assess the quality and influence of journals in their respective fields.

کارهای دیگری نظیر پیشنهاد سایر مقالات مشابه با مقاله تحت مطالعه، سایر مقالات نویسندگان، جستجو در روابط ریاضی، پیدا کردن معانی اصطلاحات تخصصی و یا اختصارات استفاده شده، روش های ریاضی استفاده شده و توضیح درباره هر یک و ... از دیگر پرسش هایی است که میتوان از کوپایلت پرسید و جواب دریافت کرد.

با همه مزایای گفته شده ذکر این نکته مهم است که در پایان کار همه این ابزارها برای تسهیل و تسریع امر پژوهش هستند و در انتهای کار این خود محقق است که میبایست با بررسی دقیق  و مطالعه مستندات از مفاهیم آن استفاده کند. این ابزار تنها تسهیل کننده تحقیق و پژوهش بوده و به هیچ عنوان نمیتواند کار محقق را برای او انجام دهد. ذکر این نکته هم خالی از لطف نیست که خود سازندگان کوپایلت همواره میگویند که پاسخ های دریافتی از هوش مصنوعی میتواند با اشکلالاتی همراه باشند و در روند استفاده از آن حتما میبایست به این نکته نیز توجه نمود. 

ثبت نظر:



وبسایت Vδ

وی‌دلتا، مجموعه‌ای از مقالات، آموزش‌ها و مباحث حوزه برق و برنامه‌نویسی. برای اطلاع از انتشار مقالات حتما در کانال تلگرام ما عضو شوید.

آدرس کانال تلگرام

Channel ID: @vdelta_ir

آخرین مطالب

درباره نویسنده

آرش صمدی

دانشجوی دکتری مهندسی برق قدرت و علاقه‌مند به مباحث حوزه برنامه‌نویسی کاربردی در صنعت برق. زمینه‌های تخصصی: حفاظت سیستم‌های قدرت، ارتینگ و توسعه نرم‌افزارهای حوزه مهندسی برق

سایر مقالات درباره نویسنده