CoDi: Conversational Distillation for Grounded Question Answering
Fuente:
arXiv
Salvato in:
| Autori principali: | , , , , , , , , |
|---|---|
| Natura: | Preprint |
| Pubblicazione: |
2024
|
| Soggetti: | |
| Accesso online: | |
| Tags: |
Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
|
| _version_ | 1866914918009143296 |
|---|---|
| author | Huber, Patrick Einolghozati, Arash Conway, Rylan Narang, Kanika Smith, Matt Nayyar, Waqar Sagar, Adithya Aly, Ahmed Shrivastava, Akshat |
| author_facet | Huber, Patrick Einolghozati, Arash Conway, Rylan Narang, Kanika Smith, Matt Nayyar, Waqar Sagar, Adithya Aly, Ahmed Shrivastava, Akshat |
| contents | Distilling conversational skills into Small Language Models (SLMs) with approximately 1 billion parameters presents significant challenges. Firstly, SLMs have limited capacity in their model parameters to learn extensive knowledge compared to larger models. Secondly, high-quality conversational datasets are often scarce, small, and domain-specific. Addressing these challenges, we introduce a novel data distillation framework named CoDi (short for Conversational Distillation, pronounced "Cody"), allowing us to synthesize large-scale, assistant-style datasets in a steerable and diverse manner. Specifically, while our framework is task agnostic at its core, we explore and evaluate the potential of CoDi on the task of conversational grounded reasoning for question answering. This is a typical on-device scenario for specialist SLMs, allowing for open-domain model responses, without requiring the model to "memorize" world knowledge in its limited weights. Our evaluations show that SLMs trained with CoDi-synthesized data achieve performance comparable to models trained on human-annotated data in standard metrics. Additionally, when using our framework to generate larger datasets from web data, our models surpass larger, instruction-tuned models in zero-shot conversational grounded reasoning tasks. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2408_11219 |
| institution | arXiv |
| publishDate | 2024 |
| record_format | arxiv |
| spellingShingle | CoDi: Conversational Distillation for Grounded Question Answering Huber, Patrick Einolghozati, Arash Conway, Rylan Narang, Kanika Smith, Matt Nayyar, Waqar Sagar, Adithya Aly, Ahmed Shrivastava, Akshat Computation and Language Artificial Intelligence Distilling conversational skills into Small Language Models (SLMs) with approximately 1 billion parameters presents significant challenges. Firstly, SLMs have limited capacity in their model parameters to learn extensive knowledge compared to larger models. Secondly, high-quality conversational datasets are often scarce, small, and domain-specific. Addressing these challenges, we introduce a novel data distillation framework named CoDi (short for Conversational Distillation, pronounced "Cody"), allowing us to synthesize large-scale, assistant-style datasets in a steerable and diverse manner. Specifically, while our framework is task agnostic at its core, we explore and evaluate the potential of CoDi on the task of conversational grounded reasoning for question answering. This is a typical on-device scenario for specialist SLMs, allowing for open-domain model responses, without requiring the model to "memorize" world knowledge in its limited weights. Our evaluations show that SLMs trained with CoDi-synthesized data achieve performance comparable to models trained on human-annotated data in standard metrics. Additionally, when using our framework to generate larger datasets from web data, our models surpass larger, instruction-tuned models in zero-shot conversational grounded reasoning tasks. |
| title | CoDi: Conversational Distillation for Grounded Question Answering |
| topic | Computation and Language Artificial Intelligence |
| url | https://arxiv.org/abs/2408.11219 |