- IPL Dataset: https://creativeag-ipl-dataverse.streamlit.app/
- Research Article Summarisation: https://paper-summarizer-gradientgeeks25.streamlit.app/
- IPL Dataset: https://github.com/Anurag-ghosh-12/ipl
- Research Article Summarization : https://github.com/Suchana4Hazra/Research_Paper_Summarizer
🚀 Exciting News! Revelation 2K25 is back, proudly presented by the Department of Computer Science and Technology, IIEST Shibpur. This year, we are thrilled to announce our collaboration with multiple leading product-based companies eager to recruit top talent from our main technical events. Don't miss this unparalleled opportunity to showcase your skills, stand out, and secure your future in the tech world! See you at Revelation 25! 🎉 #Revelation25 #TechFest #IIESTS #CareerOpportunities
Get ready for an exciting challenge in data analysis and machine learning at Revelation '25 with BrainDead! 🚀
We are excited to present two compelling problem statements for this flagship competition. Both problems are compulsory—whether you solve one, both, or even just a part, we encourage you to make a submission and showcase your expertise! On behalf of Team Revelation, we wish you all the best! 💻🔍 #BrainDead #Revelation25 #DataAnalysis #MachineLearning
📧 Email: revelationiiest@gmail.com
📲 Join WhatsApp Group: Cli
8000
ck here
- For Problem Statement 1 & 2, prepare a concise report in .pptx or .pdf format.
- Append your Problem Statement 2 report to the Problem Statement 1 report.
- Upload your
iPython notebook (.ipynb)to GitHub and share your public repository link via UnStop.
- Both problem statements are compulsory.
- You can submit whichever problem you solve, but your score for a partially solved problem will be given based on the report and result.
- For any unattempted problem, your score for that question will be 0.
- If you submit the report but not the iPython notebook, your score for that problem will be 0 (Aryabhatta to be blamed! 🙃).
Long story short, create a single concise report covering both problem statements (if you have solved both) and upload the corresponding code to GitHub.
🎯 Best of luck! Show us your skills and make your mark at Revelation 2K25! 🚀
Problem Statement 1: Statistics is All You Need: IPL Data Analysis and 2025 Winner Prediction – The Game Behind the Game!
Cricket is the most popular sport in India. There are various formats of this game and the most loved one is the Indian Premier League (IPL). This professional Twenty20 cricket league in India gets contested during March or April and May of every year by eight teams representing ten different cities on India. The league was founded by the Board of Control for Cricket in India (BCCI) in 2008. The IPL has an exclusive window in ICC Future Tours Programme. It is the most-attended cricket league in the world. Currently, it’s the 18th season of IPL.
You have to perform a comprehensive analysis of IPL data from its inception through to the most recent season in 2024, aimed at uncovering key insights, trends, and patterns. It should consists of data collection, preprocessing, and exploratory data analysis (EDA) to visualize metrics such as win rates, player performance, and team statistics. The analysis includes statistical insights to identify significant factors influencing match outcomes. You may use pandas and NumPy for data manipulation and matplotlib and seaborn for data visualization.
Also, try to develop an ensemble model, combining different classifier models (one such example is the ensembling of classifiers like Random Forest and XGBoost) for predicting the winner of the 2025 IPL season. It should explain the model’s features, training, validation, and performance evaluation. Additionally, you can explore experimenting with neural networks. The results section must present the model’s predictions for the 2025 season, and discussion regarding the potential strengths and limitations, and should provide insights into the predicted performance of teams and key players. The primary objective is to use historical IPL data to build a predictive model for future match prediction outcomes, demonstrating the application of advanced machine learning techniques to sports data.
- Ensure that no missing values or outliers exist in the dataset.
- Handle potential issues that could impact insights and predictions.
Perform the following analyses based on the IPL 2008-2024 Dataset:
- Plot Matches Played and Winning Percentages
- Plot Run Rate and Economy Rate (as a bowling side)
- Plot Highest and Lowest Scores
- Plot Total 4s and 6s
- Plot Average Powerplay and Death Overs Score
- Powerplay Analysis
- Get the top 20 run-scorers
- Plot Batting Average vs Batting Strike Rate for the top 20 run-scorers
- Find Highest Average and Strike Rate for players with >50 matches
- Plot top wicket-takers
- Plot top highest individual scores
- Man of the Match Count Analysis
- Use K-Means Clustering to plot Batting Average vs Bowling Economy Rate for number of clusters = 3 (Batsman, Bowler, All Rounder)
- Identify Top 10 Batsmen in each run category:
- Top 6’s scorer
- Top 4’s scorer
- Top 2’s scorer
- Top 1’s scorer
- Calculate average runs per match per season
- Identify targets of 200+ runs per season
- Find the average score of each team per season
- Analyze runs of Orange Cap Holders per season
- Track wickets of Purple Cap Holders per season
- Find top 10 bowlers per season
- Extract key features from
matches.csvdataset. - Extract crucial insights from
deliveries.csvdataset.
4️⃣ Winner Prediction Model: Develop a prediction model based on the above analyses to predict the winner of 2025 IPL..
Participants may use the following tools:
- MS Excel
- Tableau/Power BI
- Jupyter Notebook/Google Colab with Matplotlib
The dataset consists of two separate CSV files: matches and deliveries. These files contain the information of each match summary and ball by ball details, respectively. BrainDead IPL Complete Dataset (2008-2024)
The matches.csv file consists of the match informations of 1095 face-offs amongst the teams in all the IPL seasons in the last 17 years!
- 'id': Unique identifier for each match.
- 'city': City where the match was played.
- 'date': Date of the match.
- 'player_of_match': Name of the player who was awarded "Player of the Match."
- 'venue': Stadium or venue where the match was played.
- 'neutral_venue': Binary indicator (0 or 1) indicating if the match was played on a neutral venue (1) or not (0).
- 'team1': Name of the first team participating in the match.
- 'team2': Name of the second team participating in the match.
- 'toss_winner': Name of the team that won the toss.
- 'toss_decision': Decision taken by the toss-winning team (either 'field' or 'bat').
- 'winner': Name of the team that won the match.
- 'result': The result of the match (e.g., 'runs', 'wickets', 'tie', etc.).
- 'result_margin': The margin by which the winning team won the match (e.g., runs or wickets).
- 'eliminator': Binary indicator (0 or 1) indicating if the match was decided by an eliminator (1) or not (0).
- 'method': The method used to decide the match (e.g., Duckworth-Lewis, etc.).
- 'umpire1': Name of the first on-field umpire.
- 'umpire2': Name of the second on-field umpire.
The deliveries.csv file consists of ball by ball informations of all the 1095 face-offs. The dataset consists of 14,26,312 delivery entries and 17 attributes.
- 'match_id': Unique identifier for each match.
- 'inning': The inning number of a match
- 'batting_team': The name of the batting team
- 'bowling_team': The name of the bowling team
- 'over': The over number in the ongoing inning for the batting team
- 'batter': The name of the batsman (at the striker end)
- 'bowler': The name of the bowler
- 'non_striker': The name of the batsman at the non-striker end
- 'batsman_runs': Runs scored by the batsman (at the striker end)
- 'extra_runs': Extra runs (if any) conceded by the bowler
- 'total_runs': The total runs conceded by the bowler (including the runs scored by the batsman and the extra runs conceded)
- 'extra_type': The type of extra runs conceded (wide, no-ball, bye, leg-bye, etc.)
- 'is_wicket': A flag variable to indicate if there is a dismissal in a particular ball ('0' means 'no dismissal'; '1' means 'dismissal')
- 'player_dismissal': The name of the player (batsman) who got dismissed
- 'dismissal_kind': The type of dismissal (caught, bowled, run-out, LBW, stumping, obstructing the field, etc.)
- 'fielder': The name of the fielder (may be bowler also) participated in the dismissal
- 1 run is given as an extra run to the batting team in case of wide and no-ball.
- The bowler has to bowl one extra delivery in case of wide and no-ball.
- The runs scored by the batsman in case of bye and leg-bye will not get counted in the batsman's individual run, rather, it will get added to the team total.
- The 'super over' situation in case of a tie should not be considered in any evaluation. Those are just for tie-breaking purposes.
- In December 2018, the team changed its name from the Delhi Daredevils to the Delhi Capitals.
- Sunrisers Hyderabad replaced the Deccan Chargers in 2012 and debuted in 2013.
Batting Average = Total Runs Scored / Number of times out
Batting Strike Rate = (Total Runs / Total Balls Faced) * 100
Bowling Economy Rate = Total Runs Conceded / Total Overs Bowled
- A complete report detailing dataset analysis and methodology
- The report should be concise, including references, tables, figures, and results
- Clear methodology for prediction
- If Excel is used, mention the Excel formulas applied
- Source code for Exploratory Data Analysis (EDA) and 2025 IPL Winner prediction.
Participants will be evaluated based on:
- Writing style, references, and presentation of findings
- Dataset exploration and insights derived
- Methodologies used for analysis and prediction
- Source code
- Results and discussion
With the exponential rise in scientific publications, researchers struggle to keep up with the vast volume of literature. The challenge lies in creating a state-of-the-art model that accurately summarizes research articles while preserving key insights and readability, enabling researchers to assimilate information swiftly. Scientific papers differ significantly from general text due to their structured frmat (Introduction, Methods, Results, Discussion, etc.), citation dependencies, and inclusion of figure/tables. Summarizing such documents requires handling domain-specific challenges, maintaining semantic coherence, and ensuring faithful knowledge retention. This competition challenges participants to develop a framework for summarizing research artciels, leveraging state-of-the-art approaches, preferably by using Large Language Models (LLMs) with an aim to develop a hybrid summarization model that surpasses existing benchmark models while maintaining computational efficiency.
Participants must develop an extractive-abstractive hybrid model that:
- Summarized single and multi-document research papers
- Handles long-document summarization constraints while maintaining efficiency
Participants will be provided with multiple research article datasets, including a proprietary dataset from IIEST Shibpur, CompScholar, and two publicly available benchmark datasets (PubMed and arXiv). The task is to develop an efficient text-based abstractive summarization model that generates concise, coherent, and relevant summaries while outperforming existing models. Participants must also compare their models against state-of-the-art summarization frameworks (e.g., SummRuNNer, Pointer-Generator, Discourse Aware, BERT, T5, BART, PEGASUS, GPT-4, Longformer, and LED), providing a detailed performance analysis with citations.
- CompScholar Dataset (Dataset Link).
- PubMed Dataset (Dataset Link 1) (Dataset Link 2)—contains millions of biomedical research articles with structured abstracts.
- arXiv Dataset (Dataset Link)—includes structured documents across multiple scientific disciplines.
The datasets provided for this competition contain full-text research articles from diverse disciplines, including science, medicine, and engineering. Each dataset consists of article abstracts and full texts, with labeled summaries provided for supervised learning.
The datasets contain the following:
- Research Articles: Text files containing full-length papers and abstracts.
- Metadata: Titles, keywords, citations, author information, and journal sources.
- Summarization Labels: Ground truth summaries for evaluation and model training.
The CompScholar.csv file consists of the metadata informations instead of full-length article text for 370 research articles from a variety of domains.
- 'Paper Id': Unique identifier for each research paper
- 'Paper Title': The full title of the paper
- 'Key Word': Keyword of the paper
- 'Abstract': The full-length abstract from the research article
- 'Conclusion': The conclusion section from the research article
- 'Document': Combined previous metadata (Paper Title, Key Word, Abstract, Conclusion) and kept in the 'Document' attribute
- 'Paper Type': This particular field states the type of paper with values like Text Summarization, Natural Language Processing, Text Generation, Medical Data, Neural Network, etc.
- 'Summary': Manually annotated summary of that particular research article
- 'Topic': The topic attribute specifies the major domain under which that partcular paper falls, like Natural Language Processing, Medical Data Analysis and Deep Learning and Machine Learning
- 'OCR': Additional text information extracted from the diagrams, tables, images, output graphs using Optical Character Recognition (OCR) [May not be needed in the summarization task]
- 'labels': Each of the paper has been assigned one label based on the content and domains in which it falls, 'Deep Learning and Machine', 'Medical Data Analysis', 'Object Recognition' and Sentiment Analysis. [May not be needed in the summarization task]
The arXiv dataset consists of 1.7 million arXiv articles for applications like trend analysis, paper recommender engines, category prediction, co-citation networks, knowledge graph construction and semantic search interfaces.
{
"id": "0704.0002",
"submitter": "Louis Theran",
"authors": "Ileana Streinu and Louis Theran",
"title": "Sparsity-certifying Graph Decompositions",
"comments": "To appear in Graphs and Combinatorics",
"journal-ref": "None",
"doi": "None",
"report-no": "None",
"categories": "math.CO cs.CG",
"license": "http://arxiv.org/licenses/nonexclusive-distrib/1.0/",
"abstract": "We describe a new algorithm, the $(k,\\ell)$-pebble game with colors, and use\nit obtain a characterization of the family of $(k,\\ell)$-sparse graphs and\nalgorithmic solutions to a family of problems concerning tree decompositions of\ngraphs. Special instances of sparse graphs appear in rigidity theory and have\nreceived increased attention in recent years. In particular, our colored\npebbles generalize and strengthen the previous results of Lee and Streinu and\ngive a new proof of the Tutte-Nash-Williams characterization of arboricity. We\nalso present a new decomposition that certifies sparsity based on the\n$(k,\\ell)$-pebble game with colors. Our work also exposes connections between\npebble game algorithms and previous sparse graph algorithms by Gabow, Gabow and\nWestermann and Hendrickson.\n",
"update_date": "2008-12-13"
}- 'id': ArXiv ID (can be used to access the paper)
- 'submitter': Who submitted the paper
- 'authors': Authors of the paper
- 'title': Title of the paper
- 'comments': Additional info, such as number of pages and figures
- 'journal-ref': Information about the journal the paper was published in
- 'doi': Digital Object Identifier
- 'report-no': Report Number
- 'abstract': The abstract of the paper
- 'categories': Categories / tags in the ArXiv system
PubMed comprises more than 36 million citations for biomedical literature from MEDLINE, life science journals, and online books. Citations may include links to full-text content from PubMed Central and publisher web sites.
{
"MedlineCitation": {
"PMID": 0,
"DateCompleted": {"Year": 0, "Month": 0, "Day": 0},
"NumberOfReferences": 0,
"DateRevised": {"Year": 0, "Month": 0, "Day": 0},
"Article": {
"Abstract": {"AbstractText": "Some abstract (can be missing)" },
"ArticleTitle": "Article title",
"AuthorList": {"Author": [
{"FirstName": "John", "ForeName": "Doe", "Initials": "JD", "CollectiveName": ""}
{"CollectiveName": "The Manhattan Project", "FirstName": "", "ForeName": "", "Initials": ""}
]},
"Language": "en",
"GrantList": {
"Grant": [],
},
"PublicationTypeList": {"PublicationType": []},
},
"MedlineJournalInfo": {"Country": "France"},
"ChemicalList": {"Chemical": [{
"RegistryNumber": "XX",
"NameOfSubstance": "Methanol"
}]},
"CitationSubset": "AIM",
"MeshHeadingList": {
"MeshHeading": [],
},
},
"PubmedData": {
"ArticleIdList": {"ArticleId": "10.1002/bjs.1800650203"},
"PublicationStatus": "ppublish",
"History": {"PubMedPubDate": [{"Year": 0, "Month": 0, "Day": 0}]},
"ReferenceList": [{"Citation": "Somejournal", "CitationId": 01}],
},
}Main Fields will probably interest the participants are:
- "MedlineCitation" > "Article" > "AuthorList" > "Author"
- "MedlineCitation" > "Article" > "Abstract" > "AbstractText"
- "MedlineCitation" > "Article" > "Article Title"
- "MedlineCitation" > "ChemicalList" > "Chemical"
- "MedlineCitation" > "NumberOfReferences"
📊 Data Field Description of 'PubMed' file (Dataset 2):
- 'id': paper id
- 'article': a string containing the body of the paper
- 'abstract': a string containing the abstract of the paper
📊 Data Splits of 'PubMed' file (Dataset 2):
| Dataset Split | Number of Instances |
|---|---|
| Train | 119,924 |
| Validation | 6,633 |
| Test | 6,658 |
The competition will focus on extractive-abstractive summarization techniques. Participants will explore LLM-based models to generate meaningful summaries that maintain contextual relevance.
- A novel text summarization model designed to generate concise, coherent, and accurate summaries of research articles.
- Trained models demonstrating superior performance compared to existing benchmark summarization frameworks.
- Comparative results with state-of-the-art models, including SummRuNNer, Pointer-Generator, Discourse Aware, BERT, T5, BART, PEGASUS, Longformer, LED, GPT-4-based summarization approaches, Other models cited in recent research papers
- A comprehensive evaluation report, comparing results with existing benchmarks.
Models will be evaluated on:
- ROUGE Scores (ROUGE-1, ROUGE-2, ROUGE-L)—Measures the overlap of generated summaries with reference summaries.
- BLEU Score—Evaluates text fluency and translation quality.
- Summarization Length and Readability—Ensuring meaningful information is retained in concise form.
- Computational Efficiency—Assessing training time, inference speed, and memory usage.
Ranked Model Performance Table (for CompScholar Dataset):
| Rank | Model | ROUGE-1 | ROUGE-2 | ROUGE-L | BLEU |
|---|---|---|---|---|---|
| 1 | PEGASUS | 45.1 | 21.8 | 42.3 | 36.2 |
| 2 | BART | 43.5 | 19.4 | 40.6 | 33.8 |
| 3 | Longformer | 41.2 | 18.9 | 39.1 | 32.4 |
| 4 | LED | 40.5 | 17.8 | 38.6 | 31.7 |
| 5 | GPT-4-Summarization | 39.2 | 16.5 | 37.2 | 30.8 |
Everyone should use the mentioned environments for coding:
- Google Colab or Local System
- Jupyter Notebook
- TensorFlow, Keras, Python, PyTorch
- Your notebook name should follow this format:
team_name_brain_dead_2k25.ipynb
- Submission via Unstop.
- If there is an issue, contact us at revelationiiest@gmail.com.
- Join WhatsApp Group: Click here
- A complete report of the methodology employed in your work and the source code of your best pipelines for each selected dataset. The report should be concise. This report might include references, tables, figures, and results. Everyone should submit their code between 21/03/2025 at 10:00 AM to 23/03/2025 at 10:00 AM.
- Table of Contents
- Introduction
- Dataset Preprocessing
- Model Architecture and Training Methodology
- Performance Evaluation
- Results and Discussion
- Summary of model performance on the test set
- Optimization Strategies
- Conclusion
- References
Upload your code to Github. Create an account if you don’t have one. Make sure your repository is public. Your report should have most of these: Writing Style, references, figures, etc. Dataset exploration Methodology Results of analysis Comparative study Pipeline architecture you used Conclusion
We'll be checking all submissions for plagiarism to ensure everything's original and above board. It's all about keeping things real and honest! 🚀
🚀 Best of luck! Show us your analytical and predictive skills!
CricSheet (IPL Dataset)
- U.S. National Library of Medicine (PubMed Dataset)
- Clement, C.B., Bierbaum, M., O'Keeffe, K.P. and Alemi, A.A., 2019. On the use of arxiv as a dataset. arXiv preprint arXiv:1905.00075. (arXiv Dataset)