Skip to content

LINGO LABSIIT Gandhinagar

Research

Themes

Work in multilingual and multimodal intelligence, foundation model architectures, reasoning and agentic systems, and evaluation and trustworthy AI.

Theme

Multilingual & Multimodal Intelligence

Language diversity, code-mixing, and multimodal understanding.

Projects

Funded project · ANRF

Code-mixed NLP

Hindi–English benchmarks and tools, 2023–2026.

Page → Funded project · ANRF

Chitrabhasha

Multimodal datasets and foundation models, 2026–2029.

Resources and papers

Paper

TCFM

Task-conditional multilingual embedding adaptation.

Page → Dataset

IndicTalk

Multilingual conversational corpus.

Page → Dataset

Triveni

Indic multimodal captions.

Page → Dataset

MUTANT

Multi-sentential Hinglish.

Page → Dataset

MMT

Indian social-media corpus.

Page → Dataset

HinGE

Hinglish generation and evaluation.

Page → Dataset

PHINC

Parallel Hinglish for MT.

Page →

Theme

Foundation Model Architectures & Efficient Intelligence

New architectures, efficient computation, and model adaptation.

Projects

Funded project · IndiaAI / MeitY

Project EKA

India-centric foundation LLM programme.

Page → Project

Project Unity

Ganga: open Hindi language models.

Resources and papers

Tokenizers

Ansh

Indic tokenizers (128k–256k).

Page →

Theme

Reasoning, World Models & Agentic Intelligence

Reasoning, procedural execution, world models, and grounded intelligence.

Resources and papers

Paper

TempUN

Numerical-temporal reasoning in LLMs.

Page →

Theme

Evaluation, Reliability & Trustworthy AI

Evaluation, alignment, robustness, security, and interpretability.

Projects

Funded project · IndiaAI / MeitY

Project EKA

Eka-Eval: Indian-language LLM evaluation suite.

Page → Project

Project Unity

UnityAI-Guard: multilingual toxicity detection for Indian languages.

Resources and papers

Benchmark

PythonSaga

Code-generation evaluation.

Page → Benchmark

LEGOBench

Scientific leaderboard generation.

Page → Dataset

PI-Indic-Align

Persona–instruction alignment.

Page → Paper

X-BAT

Cross-lingual backdoor attacks.

Page → Paper

XME

Cross-lingual model editing.

Page → Dataset

SangrahaTox

Multimodal cultural toxicity.

Page →

Past

Past programmes

NLPExplorer, Gandhipedia, and scholarly document AI.

Projects

NLPExplorer

Portal for NLP papers, authors, and venues.

Page → Funded project · NCSM

Gandhipedia

Portal for Gandhian literature and networks.

Page → Funded project · SERB

Scholarly document AI

Table and chart extraction, 2019–2022.

Resources and papers

System

OCR++

Scholarly PDF extraction.

Page → Dataset

COMPARE

Peer-review comparisons.

Page → Paper

CoSAEmb

Scientific-article embeddings.

Page → Dataset

Gurukul

Textbook question–answer pairs.

Page → Paper

Model Hubs

Popularity vs performance.

Page → Paper

Multi-Annotation

Annotator disagreement.

Page → Dataset

PoliWAM

2019 election WhatsApp corpus.

Page →