IEEE NU CIS

BETA

WHO WE ARE

IEEENU CIS is a student chapter that was founded near the end of 2021 as the second student chapter from the IEEENU Student Branch. It was a club founded on the premise of spreading knowledge and awareness about computational intelligence. Computational intelligence covers a lot of fields under its umbrella. With topics ranging from fuzzy logic to genetic algorithms to evolutionary programming to machine learning and deep learning, computational intelligence is a vast category of science that can be summarized under the title of “creating systems that allow computers to think intelligently and make smart decisions” In IEEENU CIS, we focus on preparing university undergraduates to be prepared for the job market. Let’s walk through that process.The job market for AI requires skills that can be categorized into 3 types

  • Dealing with Data
  • Knowledge of the Practice
  • Data Visualization

Dealing with Data:

job candidate should be able to handle the input, retrieval, cleaning, and wrangling of data into a usable format for the next stage of modeling. i.e. SQL, data engineering, data exploration, etc.

Knowledge of the Practice:

job candidate should be able to create, formulate, and tune systems involving computational intelligence i.e. ML and DL models, fuzzy logic controllers, expert systems, etc

Data Visualization:

job candidate should be able to properly visualize and communicate his findings and results to non-technical stakeholders.

Members of the club are exposed to activities that aim to improve one or more of these aspects. In the beginning, members are exposed to a training period of 4-5 weeks which aims to teach them about the basic building blocks of machine learning and deep learning as well as get them familiar to basic coding. The courses used as studying material are handpicked from trusted sources like Udacity and also from prestigious universities and by talented instructors to provide the most up to date and strong theoretical knowledge that members will need in practice.New members are also assisted by senior members (who are labeled trainers) to better the learning experience.The training cycle includes video material, text material, quizzes, coding questions and mentoring. Once the training cycle ends, members enter the competition cycle where they gain access to 3 distinct weekly activities:

Members enter the competition cycle where they gain access to 3 distinct weekly activities:

  • Kaggle Competitions
  • Variety Tournaments
  • Problem Solving

And they also gain access to committees after 2 or 3 weeks in the competition cycle. The available committees are:

  • Trainers
  • Kagglers
  • Variety
  • Opportunities
  • Content Creation
  • Problem Solving

Club Activities

As a general rule in the club, during every week, every member has the full freedom to choose whichever activities he/she would like to participate in. This allows full flexibility for members to be able to maintain consistency of general participation in club activities rather than getting burned out. This allows them to be able to pick and choose the activities for the week that best suit their time schedule.

However, each activity in the club adds points to the participating member's score. These points are then what represent this member's activity/contribution to the club, and these points are what decide who gets the end-of-year rewards and who gets priority in general for different opportunities/jobs that the club can recommend for.

Kaggle Competitions

Kaggle Competitions are the bread and butter of the club's practice. Each week, a new dataset is presented to members as they compete against each other to gain the highest score on the leaderboard. After the competition, members are given a score based on their overall score (accuracy, MSE, MAE, F Score, etc.) compared to others, their data exploration, the readability of their work, and the inference time of their models.

This particular activity boosts members' ability to deal with different types of data, as members are asked to explore their targeted data and learn how to deal with different data types ranging from images and audio to video and language. It is our main promoter for learning the practice as members put their learned concepts to work and attempt to implement the latest models and technologies to get the highest scores. It also boosts data visualization skills as members need to properly organize their work and present it in an acceptable manner and properly visualize their data and modeling.

This activity starts as individuals, but as members get more adept and familiar with the system, they are organized into teams for a multitude of reasons:

  • Provide better results overall as a team
  • Teach teamwork, leadership, and build trust between team members
  • Prepare members for the usual formats of hackathons which is team-based

Variety Tournaments

In the weekend of every week, we host a variety event. Variety events include any coding activity/game that is not our usual kaggle competition. These activities promote adaptability, problem-solving, and competitiveness. There is not one particular activity but for the sake of example, Battlesnake is explained here:

Battlesnake tournaments are tournaments conducted on the online battlesnake platform as members attempt to build the "smartest" snake to win it all. Snakes are basically web servers that act as controllers to snakes in the game. This activity promotes a lot of creativity as players can use whatever approach they think of to make their snake act more intelligently and dodge more pitfalls. Anything from pathfinding to mapping algorithms to ML or even reinforcement learning or RL is fair game, and it makes for some very intense games of the once very simple game of snake. This activity promotes dealing with data as members have to deal with an environment-based problem instead of a dataset-based one. They learn how to receive inputs and process them to make logical decisions; in addition, every solution approach may involve certain transformations to the data which members learn along their journey through the leaderboards. It also promotes knowledge of the practice as members learn different approaches people have come up with in the past and also make up their own flavors of approaches to solve this deceivingly simple game.

Problem Solving

During every week, a codeforces problem sheet is prepared by our Problem Solving Committee and made available to our members. This is not a problem-solving club, so this sheet is not meant to teach members the full spectrum of problem-solving questions but rather get them familiar with questions that are most likely to be asked in job interviews. We have handpicked the topics that show up in interviews the most, and we even add a bit of flavor to the questions by trying, as much as we can, to make them related to AI engineering problems.

The sheet has ramping difficulty so as to pose a challenge for some members and be a fun weekly puzzle for others who just want to get familiar with problem-solving basics.

Club Committees

After a 2-3 weeks of the competition cycle, members can enter one of 6 committees to help give back to the club and keep the wheel running. These committees are the Trainers, Kagglers, Variety, Opportunities, Content Creation, and the Problem Solving committees. From their names, most of these committees aim to give senior members the chance to help newer members go through the process they did and, in doing so, improve the system and the learning environment.

Trainers are responsible for the main training period in the beginning and also for supervising any course content instructed in the club. After that, in the competition cycle, Kagglers and Variety are responsible for helping new members find their way through the beginner Kaggle competitions and Variety tournaments by providing tips, tricks, and roadmaps of ideas for members to explore. This helps make every new batch better than the last and further supports the growth of the club. Opportunities mainly focus on finding online opportunities for club members, ranging from courses to bootcamp opportunities, Kaggle competitions of interest, Variety events, hackathon opportunities, or even just project ideas that members can work on. Content Creation includes members who are interested in helping develop and maintain the website, as well as in editing and preparing video content recorded by our members. Finally, the Problem Solving committee is responsible for preparing the weekly problem sheets mentioned previously and also contributes to the Problem Solving for AI playlist.

Through this journey, in addition to the main 3 job-related goals we envision, we also target many other interpersonal and soft skills that members pick up during their time in the club. These include but are not limited to problem-solving, leadership, teamwork, familiarity with Kaggle and data science platforms, and teaching and learning to instruct others. Members who enter committees get the chance to teach and share what they have learned with newer members, which helps them develop their ability to instruct and teach others. This can be a major contributor to their personality if they seek to pursue academia in the future or enhance their general ability to work within teams.

At the end of the year, members who are at the top of the score leaderboard are presented with job recommendations for various AI jobs. These recommendations come from the club chair and from previous club alumni who have graduated or are currently employed. This cycle allows the club to continually grow as we accumulate more alumni and, in turn, offer more recommendations, supporting the growth of more successful alumni and creating a positive feedback loop.