๐ฅ 6x Linkedln Top Voice | AI Research Scientist & Chief Data Scientist at IBM | Generative AI Expert | Author - Hands-on Time Series Analytics with Python | IBM Quantum ML Certified | 11+ Years in AI | MLOps | IIMA |
๐๐ฎ๐-๐ฏ๐ฑ๐ฏ ๐๐ผ๐บ๐ฝ๐๐๐ฒ๐ฟ ๐ฉ๐ถ๐๐ถ๐ผ๐ป ๐๐ฒ๐ฎ๐ฟ๐ป๐ถ๐ป๐ด Meta AI Introduces A New AI Technology Called โFew-Shot Learner (FSL)โ To Tackle Harmful Content Follow me for a similar post: ๐ฎ๐ณ Ashish Patel ๐ฎ๐ณ ------------------------------------------------------------------- ๐๐ป๐๐ฒ๐ฟ๐ฒ๐๐๐ถ๐ป๐ด ๐๐ฎ๐ฐ๐๐ : ๐ธ Paper: ๐๐ป๐๐ฎ๐ถ๐น๐บ๐ฒ๐ป๐ ๐ฎ๐ ๐๐ฒ๐-๐ฆ๐ต๐ผ๐ ๐๐ฒ๐ฎ๐ฟ๐ป๐ฒ๐ฟ ๐ธ This paper is published arxiv2021. ๐ธ For the training of AI models, a massive number of labeled data points or examples are required. Typically, the number of samples needed is tens of thousands to millions. Collection and labeling of these data can take several months. This manual collection and labeling delay the deployment of AI systems that can detect new types of harmful content over different social media platforms. To handle this issue, Meta has deployed a relatively new AI model called โFew-Shot Learnerโ (FSL) such that harmful contents can be detected even if enough labeled data is not available. ------------------------------------------------------------------- ๐๐ ๐ฃ๐ข๐ฅ๐ง๐๐ก๐๐ ๐ธ Large pre-trained language models (LMs) have demonstrated remarkable ability as few-shot learners. ๐นHowever, their success hinges largely on scaling model parameters to a degree that makes it challenging to train and serve. ๐ธIn this paper, we propose a new approach, named as EFL, that can turn small LMs into better few-shot learners. The key idea of this approach is to reformulate potential NLP task into an entailment one, and then fine-tune the model with as little as 8 examples. ๐นWe further demonstrate our proposed method can be: (i) naturally combined with an unsupervised contrastive learning-based data augmentation method; (ii) easily extended to multilingual few-shot learning. ๐ธA systematic evaluation on 18 standard NLP tasks demonstrates that this approach improves the various existing SOTA few-shot learning methods by 12\%, and yields competitive few-shot performance with 500 times larger models, such as GPT-3. ------------------------------------------------------------------- #computervision #artificialintelligence #innovation -------------------------------------------------------------------
๐ฅ 6x Linkedln Top Voice | AI Research Scientist & Chief Data Scientist at IBM | Generative AI Expert | Author - Hands-on Time Series Analytics with Python | IBM Quantum ML Certified | 11+ Years in AI | MLOps | IIMA |
2yPaper: https://arxiv.org/pdf/2104.14690.pdf Reference: https://ai.facebook.com/blog/harmful-content-can-evolve-quickly-our-new-ai-system-adapts-to-tackle-it/