OpenTalks.AI - Александр Артеменков, Towards Scalable Dimensionality Reduction via Noise Contrastive Estimation

Transcript

1. Towards Scalable Dimensionality Reduction Aleksandr Artemenkov joint work with Maxim Panov via Noise Contrastive Estimation
2. News Headlines Of India 3 million topics you are unfamiliar with Kaggle Dataset
4. Just repeat 3 million times! (not a good idea)
5. Is there a better approach? (visualization)
7. Noise Contrastive Estimation (without a single formula) (check out the poster for them)
8. Building Representation “cold” “frosty” “cold” “warm” “cold” “first” similar different unrelated (noisy) ? ? ?
9. Simple Insight • similar points = closer than predefined noisy distance • different points = further than predefined noisy distance Any benefits from such approach?
10. Parallelized Batch Optimization • noisy distance is predefined • repel different points • attract similar points • adjustments are independent of each other to distinguish them from noise Meet NCVis! (Noise Contrastive Visualization) GitHub Repository
11. Why use NCVis? (there are reasons) Paper
14. Thank you for your attention! (find me near the poster)