Spaces:
No application file
No application file
Update README.md
Browse files
README.md
CHANGED
|
@@ -1,10 +1,59 @@
|
|
| 1 |
---
|
| 2 |
-
title:
|
| 3 |
-
emoji:
|
| 4 |
-
colorFrom:
|
| 5 |
-
colorTo:
|
| 6 |
-
sdk: static
|
| 7 |
pinned: false
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 8 |
---
|
| 9 |
|
| 10 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
---
|
| 2 |
+
title: "MWire Labs"
|
| 3 |
+
emoji: "📡"
|
| 4 |
+
colorFrom: "yellow"
|
| 5 |
+
colorTo: "gray"
|
| 6 |
+
sdk: "static"
|
| 7 |
pinned: false
|
| 8 |
+
license: "apache-2.0"
|
| 9 |
+
tags:
|
| 10 |
+
- AI
|
| 11 |
+
- NLP
|
| 12 |
+
- Multilingual
|
| 13 |
+
- Low-resource
|
| 14 |
+
- Khasi
|
| 15 |
+
- Northeast-India
|
| 16 |
+
- Governance
|
| 17 |
+
- Civic-Tech
|
| 18 |
---
|
| 19 |
|
| 20 |
+
# MWire Labs
|
| 21 |
+
|
| 22 |
+
MWire Labs is the applied AI research wing of MWire (Formerly known as Marketing Wire).
|
| 23 |
+
We build scalable, ethical AI systems for underrepresented languages and grassroots data — pioneering research from Northeast India with global relevance.
|
| 24 |
+
|
| 25 |
+
Our work bridges the gap between **local challenges and frontier AI**, creating models and datasets that empower governance, nonprofits, and rural innovation.
|
| 26 |
+
|
| 27 |
+
---
|
| 28 |
+
|
| 29 |
+
## Focus
|
| 30 |
+
|
| 31 |
+
- **Low-Resource NLP**: Building language models for Khasi and other underserved languages.
|
| 32 |
+
- **Grassroots Datasets**: Structuring data from villages, agriculture, and civic institutions.
|
| 33 |
+
- **Applied Civic AI**: Turning AI research into practical tools for governance and entrepreneurship.
|
| 34 |
+
|
| 35 |
+
---
|
| 36 |
+
|
| 37 |
+
## Flagship Work
|
| 38 |
+
|
| 39 |
+
- [KhasiBERT](https://huggingface.co/MWirelabs/khasibert) — the first Khasi language transformer model.
|
| 40 |
+
- [Khasi-English Semantic Search](https://huggingface.co/MWirelabs/khasi-english-semantic-search) — cross-lingual search for low-resource contexts.
|
| 41 |
+
- [Northeast India Districts & Villages Dataset](https://huggingface.co/datasets/MWirelabs/Northeast-India-Districts-and-Villages) — structured registry of 15,000+ villages and 100+ districts.
|
| 42 |
+
|
| 43 |
+
---
|
| 44 |
+
|
| 45 |
+
## Roadmap
|
| 46 |
+
|
| 47 |
+
- **NeoDAC Models** — lightweight, domain-aligned LLMs for governance and civic tech.
|
| 48 |
+
- **Expanded Multilingual Embeddings** — covering more Northeast Indian languages.
|
| 49 |
+
- **Open Agriculture & Rural Datasets** — unlocking insights for farmers, SHGs, and entrepreneurs.
|
| 50 |
+
- **Applied AI Pilots** — in partnership with nonprofits and government bodies.
|
| 51 |
+
|
| 52 |
+
---
|
| 53 |
+
|
| 54 |
+
## Why It Matters
|
| 55 |
+
|
| 56 |
+
Most AI research overlooks low-resource languages and rural contexts.
|
| 57 |
+
MWire Labs is proving that **AI can be built for the margins, not just the mainstream** — starting from the Northeast, and scaling to India and beyond.
|
| 58 |
+
|
| 59 |
+
---
|