Working Group Shortlisted Proposals
Below are details of the four Working Groups (WGs) open for membership applications. Please read the descriptions below, and if you find a WG that interests you, use the link provided below to apply for participation. Please note that WG members are expected to actively contribute in accordance with the expectations and responsibilities outlined by the respective WG leaders.
In-person attendance at the conference is mandatory for all Working Group (WG) participants. Although the conference officially begins on December 3, Day 0 (December 2) is reserved exclusively for WG participants. On this day, the group will meet for a full-day dedicated to advancing the project. The WG may continue its collaborative work throughout the conference as needed.
In addition, the conference program will include a session on Day 2 (December 4) or Day 3 (December 5) during which each WG will deliver a brief presentation outlining its project and progress to conference attendees.
Please scroll further down for more information on the complete process. If you have any questions about a WG, please contact the respective WG leaders via email.
Application Link: https://forms.gle/gKQPRQvZqBTyYC737
Application Deadline: 20th June 2026 AoE
Understanding Refute Problem Difficulty for Learners in Indian Computing Education
Leader: Viraj Kumar (viraj.kumar.cs@gmail.com)
Abstract: Refute Problems present students with buggy code for a given task and ask students to demonstrate why the code is buggy. Little is known about what makes a Refute Problem difficult for students. This Working Group (WG) investigates possible causes of difficulty such as
- Comprehending the given implementation (code-level difficulty)
- Comprehending the given problem specification (task-level difficulty)
We hypothesize that task-level difficulty is particularly relevant in the context of Indian computing education, where several students may have limited fluency in the language in which the task is expressed (e.g., English). This WG will limit the scope to Refute Problems in the context of Introductory Programming (CS1), where they were first proposed and have been most extensively studied. Drawing on cognitive load theory, code comprehension, and assessment design, as well as an analysis of existing Refute Problems, instructor reflections, and exploratory studies with undergraduate learners from participating Indian institutions, this WG aims to develop a framework for characterizing Refute Problem difficulty together with design guidelines, annotated examples, and tools to help instructors tailor the difficulty of Refute Problems for their students. This will support the creation of more equitable, scalable, and cognitively calibrated Refute Problems for assessments.
Detailed Description: Detailed description can be downloaded at WG1_detailed.pdf
Assessment Integrity in Indian CS Education: GenAI-Era Capabilities, Traditional Examinations, and OBE Accreditation
Leader:Samridhi Singhal (samridhi.singhal@iilm.edu)
Abstract: The rapid adoption of Generative AI tools among the students has created a situation that Indian undergraduate Computer Science Education is structurally unprepared to handle, Students can now use tools such as ChatGPT and GitHub Copilot to generate the examination responses for almost all conventional assessment types like lab programs, theory answers, assignment submissions and even Viva preparation. Faculty at most Indian Institutions do not have the authority to modify the examination and laboratory formats prescribed by their affiliating universities. At the same time, institutions are required to demonstrate NAAC and NBA accreditation compliance by computing and reporting Course Outcome attainment percentages mapped to Program Outcomes. When the assessments underlying these figures may be susceptible to Gen-AI use, the reliability of the attainment data built on them becomes an open and important question – one that warrants rigorous empirical investigation rather than assumption. This three-way intersection that is – Gen-AI enabled student behaviour, rigid University mandated assessment formats and OBE based accreditation requirements is a distinctly Indian Education structural problem. Global computing education research on GenAI and academic integrity has not been designed for a context where faculty have limited reformative authority and where accreditation frameworks create strong institutional incentives around attainment reporting. Existing literature has not, to our knowledge, exclaimed this three-way structural challenge in the Indian Computing education context. Faculty are navigating it alone, students receive contradictory signals about what constitutes honest work, and accreditation bodies may be receiving documentation whose reliability warrants scrutiny. This Working Group sets out to empirically investigate the nature and extent of this problem.
Detailed Description: Detailed description can be downloaded at WG2_detailed.pdf
Foundations at Risk: Investigating Core Programming Competency Among Indian CS Undergraduates in the Age of AI/ML Hype and Generative AI Tools
Leader: Preeti Mehta (preeti.mehta@iilm.edu)
Abstract: There is a worrying trend across Indian undergraduate CS programmes. Students are confidently using ChatGPT to write code, but are unable to debug a simple loop, reason about a recursive function, or trace pointer errors in C. Three intersecting pressures seem to be contributing to this shift: (a) a placement-driven culture that values credentials over deep mastery; (b) AI/ML hype that encourages students to skip foundational courses in favour of advanced electives; and (c) increasing reliance on Generative AI tools that allow code generation without real understanding. While faculty concern about this issue is widespread across India, the CS education literature has paid virtually no systematic empirical attention to this phenomenon. This working group seeks to help fill that gap by undertaking an open, evidence-based exploration of whether and to what extent foundational competencies are changing, why they are changing, and how institutions might respond.
Detailed Description: Detailed description can be downloaded at WG3_detailed.pdf
Local-Language Reasoning Small Language Models for Computer Science Education in Low Resource Languages
Leader: Vikram Vincent (vikram.vincent@freethoughtlabs.com)
Abstract: Low resource languages such as Kannada, Tamil, Telugu, Malayalam, and Hindi remain under-resourced for NLP despite being spoken by millions and having rich, morphologically complex structures. Recent work highlights persistent data sparsity, script and grammar complexity, and the need for dedicated resources and models tailored to these languages. Studies on low-resource sentiment analysis in Tamil and Tulu, for instance, show that transformer-based multilingual models significantly outperform traditional ML baselines but still suffer from limited annotated data and language-specific nuances. Concurrently, national initiatives (e.g., the National Language Translation Mission and NEP) have motivated the translation of educational content. However, these translation-centric initiatives typically focus on static materials like lectures or documents. There remains a critical gap at the intersection of: (a) Edge-deployable Small Language Models (SLMs) (1–7B parameters), (b) Dravidian and other Indian languages, (c) Code reasoning tasks (e.g., programming contest problems, debugging, step-by-step logic), To address this gap, this working group will build upon ongoing experiments in foundational ML tools and local-language code reasoning. We will focus on the rigorous dataset creation and model training phases, ensuring robust pedagogical and technical evaluation before widespread curriculum integration.
Detailed Description: Detailed description can be downloaded at WG4_detailed.pdf
A Working Group (WG) is a collaborative research project conducted by a team of typically five to ten researchers from India to produce a high-value output on a topic of interest in computing education. Membership or leadership in a WG is a nearly 10-month commitment. The weekly workload may vary depending on the project. WG membership is highly competitive, and not everyone who applies will secure a spot, despite the strength of their background and credentials. The WG operates in three phases as follows:
- Before COMPUTE 2026
- During COMPUTE 2026
- After COMPUTE 2026
Before the Conference (COMPUTE 2026):
A WG begins with a proposal written by up to three WG leaders representing at least two institutions.
The WG chairs review proposals and, based on multiple criteria, decide which proposals can proceed to member recruitment. The selected WGs are posted on the ACM Compute website, and recruitment for the groups is via a form. All potential members must submit an application through the official membership recruitment form on the website. Each participant can apply to only one WG proposal. Late applications will not be accepted.
Participant recruitment for WGs occurs in two rounds. Once the first round of recruitment applications closes, the WG chairs send each WG leader a list of applicants for their group. The WG leaders select their group members from it and notify the WG chairs of their decision. There is a brief negotiation period during which the WG chairs work to resolve any oversubscribed or undersubscribed groups. At this point, WGs can begin work on their project with the members whose applications they have accepted.
If any of the approved WGs need additional members or are still seeking members with specific expertise, a second recruitment round of approximately two weeks follows. The second round is open to new applicants and applicants not approved during the first round who want to apply to another WG. The second round of recruitment is not guaranteed. The final WG comprises five to ten researchers, including the WG leaders. The viable groups (those that have recruited sufficient members) proceed with collaboration and work until the conference.
The WG will submit a 2-page extended abstract in early October (plus an optional additional page for references). Per ACM authorship rules, the abstract to be published in the conference proceedings should be something that the whole group, without exception, has contributed to.
During the Conference (COMPUTE 2026):
In-person attendance at the conference is mandatory for all Working Group (WG) participants. Although the conference officially begins on December 3, Day 0 (December 2) will be reserved exclusively for WG participants. On this day, the group will meet for a full-day dedicated to advancing the project.
The WG may continue its collaborative work throughout the conference as needed.
In addition, the conference program will include a session on Day 2 (December 4) or Day 3 (December 5) during which each WG will deliver a brief presentation outlining its project and progress to conference attendees.
After the Conference (COMPUTE 2026):
Following the conference, the WG chairs will provide general feedback on the overall direction and scope of the work (the initial go/no go decision). Upon receiving the chairs’ feedback, the WGs have some time (till March 2027) to continue their work and update the report, which is expected to then be submitted for rigorous external reviews at journals like ACM TOCE or similar.
The list below summarises the requirements for a Working Group proposal. Carefully follow the outline and include all required information. Incomplete proposals will likely be rejected.
- The proposal uses the same format as used for paper submissions
- The proposal begins with a two-page ‘paper’ with these topics
- background and related work
- goals of the proposal
- proposed methodology
- expected deliverables
- whether or not the research needs human subjects
- references (which can exceed the 2-page limit)
- The proposal also contains a two-page appendix that contains
- An outline of the inclusion criteria for member selection
- a detailed proposed work-plan with meeting frequency and milestones
- How the WG plans to incorporate members working in different time zones
- estimation of good WG size for the project
- for each of the proposers/leaders, a discussion of their prior experience in the Working Group as a leader or member or their research profile
Proposal Body (2 pages + references):
y should be used to explain the work background, goals, proposed methodology, and expected deliverables. This can, and should, read like the beginning of a research paper, as the working group should be conducting research.
WG proposals should move from “global ideas” to India-grounded, researchable problems with clear scope, population, and methodology. WG proposals should avoid broad, globally generic themes and instead focus on well-scoped, contextually grounded problems within the Indian computing education ecosystem. Proposals must clearly articulate:
- The Indian context or problem setting
- It could be computing education related to K-12, Higher Education, Distance education, Work-integrated learning, teacher training.
- The target population
- Data sources or access feasibility
- A methodology that can realistically be executed within 10 months
A successful WG proposal will draw on existing literature and outline meaningful work that warrants a working group effort by a multicultural team. In addition, it should be clear that the methodology is appropriate and that the expected results are both reasonable (achievable within 10 months) and meaningful. The same topic must not have been proposed for any WG in parallel.
Note: The successful proposal can result in two publications: 1) an extended abstract published with COMPUTE Proceedings, and 2) the full final report published after the conference in a journal. These two publications should have two different titles (reserve your preferred title for the full report). All accepted members are expected to contribute to the shape and direction of the working group and contribute until the submission of the journal paper, i.e. March 2027.
Proposal Appendix (2Pages)
The proposal appendix has a maximum length of two pages and should include the following details:
- Inclusion criteria for member selection, and how the leaders will encourage diversity of membership. The goal of WGs is to create new research collaboration opportunities, especially for people who are not yet well-connected within our community. We encourage WGs to consider at least one early-career academic/doctoral student / member with no previous WG experience.
- Schedule of expected meetings prior to the conference and milestones or other progress points. Leaders are asked to report the WG progress to WG chairs on a monthly basis.
- An estimation of the number of members that can contribute meaningfully to the project. The above-mentioned limit of 5-10 members can be flexible for justified reasons, and the group can also be larger if conditions on the ground allow.
- A discussion of proposers’/leaders’ prior WG experience.
- Recruitment ad. In the case that the WG proposal is approved, and in order to proceed to member recruitment in the fastest time possible, include the advertisement for your WG that will be published on the conference member recruitment page.
The working group process spans about 11 months, including time before and after the conference. The ‘phases’ of work include the following:
- Proposal – The proposal process is described above.
- Call for participation – Accepted proposals will be published on the conference site.
- Recruitment – Leaders of accepted working groups and WG Co-Chairs advertise to recruit members.
- Membership – WG co-chairs set up a web form for applicants to apply to WGs. In coordination with WG co-chairs, the leaders of each WG select members from the applicant pool in two phases based on their inclusion criteria. WG co-chairs monitor this process. If there is room in a given WG after the first phase, that WG can complete the group by also inviting applicants who didn’t fit in their first choice.
- Pre-conference work – Working groups work during the months prior to the conference according to the work plan submitted with the WG proposal.
- Conference presentation – Working groups present their progress to conference attendees during one of the sessions. The extended abstract is published in the COMPUTE 2026 proceedings.
- Post-Conference feedback – WG co-chairs provide feedback and approval to continue.
Post-Conference work – Working groups continue work to complete the report for about 3 months and submit the final version to ACM TOCE or a similar journal. All WG members are responsible for continuing their contributions to the paper after the conference. Leaders certify which members are eligible for authorship according to the ACM authorship policy before the paper is submitted to review.
- 1st May 2026 (Fri): Call for WG proposals ✅
22nd May 2026 (Fri)31st May 2026 (Sun): WG Proposals due ✅29th May 2026 (Fri)7th June 2026 (Sun): Notification of which WGs proceed ✅29th May 2026 (Fri)12th June 2026 (Fri): WG member applications open ✅- 20th June 2026 (Sat): WG member applications close
- 26th June 2026 (Fri): WG Notification 1, Open round 2
- 3rd July 2026 (Fri): WG Notification 2, Working group begins work
- 4th July to 4th Nov: WG Continue to meet weekly or bi-weekly and progress on the work
- 6th Oct 2026 (Fri): WG 4-page extended abstract camera-ready due (this will be published in COMPUTE proceedings).
- 2nd Dec 2026 (Wed): Full day meetings/discussion at the conference venue to advance the project
- 3rd to 5th Dec 2026 (Sat): Attend COMPUTE, continue collaboration, and present the progress to the attendees
- 6th Dec 2026 to 19th March 2027 (Fri): WG continue to work
- Polish and submit the journal manuscript to ACM TOCE (30th March 2027, tentative).
Please submit your applications to join a WG via this link before 20th June 2026 AoE.