This shared task aims to benchmark Grammatical Error Detection (GED) and Grammatical Error Correction (GEC) systems for Indian languages. This task focuses on sentence-level categorization, detecting and correcting grammatically wrong sentences.

---

Grammar Error Correction

The participants are expected to generate the grammatically correct version of the input sentence, preserving the original intent and structure as much as possible. (Sequence Generation)

---

Evaluation Metrics

• GLEU Score

GitHub Link

Further details and submission links can be found on GitHub.

• https://github.com/BHASHA-Workshop/IndicGEC2025/

This shared task aims to introduce and regularize the concept of Word Grouping in Indian languages. Word grouping is a process wherein, given a plain text sentence in an Indian language, the task is to output a sequence of word groups, where each group represents a semantically cohesive unit.

---

Supervised Monolingual Word Group Identification

The participants are expected to build models for word group identification within a single Indian language, using fully supervised data. Training and development data will be provided for Hindi language. Participants must predict semantically cohesive word groups on a held-out test set in the same language. Use of any approach -- rule-based, statistical, or neural models -- is allowed. Evaluation will focus on boundary-level matching and group-level correctness. This task serves as the foundational task to establish baselines for word grouping quality.

---

Evaluation Metrics

• Exact Match F1-Score

GitHub Link

Further details and submission links can be found on GitHub.

• https://github.com/BHASHA-Workshop/IndicWG2025/