Data Science fwdays'19
2019-09-07
Kyiv, Ukraine
| title | author | length |
|---|---|---|
| Explaining a Machine Learning blackbox | Oleksander Krakovetskyi | 46m |
Data Science fwdays'19 conference took place on September 7 in Kyiv:
https://fwdays.com/event/data-science-fwdays-2019
Presentation… | ||
| Shadows Generation in the Wild | Taras Lehinevych | 38m |
Data Science fwdays'19 conference took place on September 7 in Kyiv:
https://fwdays.com/event/data-science-fwdays-2019
Presentation… | ||
| Feature engineering for abusive language detection for the Ukr language | Valeriia Lakusta | 33m |
Data Science fwdays'19 conference took place on September 7 in Kyiv:
https://fwdays.com/event/data-science-fwdays-2019
Presentation:
https://www.slideshare.ne… | ||
| Practical Few-Shot Learning | Kyryl Truskovskyi | 38m |
Presents a practical few‑shot learning recipe that often outperforms complex meta‑learning: fine‑tuning with normalized weight vectors plus cosine scoring… | ||
| Applying deployment oriented mindset for building Machine Learning models | Marianna Diachuk | 39m |
Claims hands‑on experience building two deployment services and improving their architecture, with concrete tools and fixes: use of SageMaker… | ||
| WaveGlow: Generative Modeling for Audio Synthesis | Taras Sereda | 47m |
Waveglow is a flow-based vocoder that converts mel-spectrograms into raw waveforms by using invertible affine coupling layers and 1×1 invertible convolutions… | ||
| Scaling Data Science | Denys Kovalenko | 38m |
Bolt builds a centralized feature store with SQL/JSON feature definitions, daily snapshots and data‑quality checks, plus Redis for low‑latency lookups, Elastic… | ||
| OCR in the Wild World of Social Media | Evgen Terpil | 41m |
An OCR pipeline for noisy social-media images that detects text regions, crops them, and runs recognition using SSD-style detectors with a VGG16 encoder and… | ||
| Programmatically creating and managing training data | Braden Hancock | 36m |
Programmatic labeling replaces per-example annotation with small programs called labeling functions that output labels or abstain and a label model learns… | ||
| Data-Driven Solutions in Engineering | Vladislav Rosov | 34m |
Presents data-driven engineering solutions using neural-network surrogates and physics-informed models. Describes a virtual twin for a vibrating wind turbine… | ||
| How We Built a Safety Violations Detection System | Yehor Morylov | 38m |
A safety-violation detection system for video streams combines a modified Faster R-CNN detector with object tracking and Kalman interpolation so the detector… | ||
| Transfer learning in NLP | Thomas Wolf | 44m |
Transfer learning uses large pretraining plus fast fine‑tuning and delivers state‑of‑the‑art gains: masked language modeling and autoregressive LM objectives… | ||
| Highlights | 03m | |
Conference highlight reel that praises organization and crowd and offers impressions rather than technical content. Mentions deploy a model to production and… | ||
| Panel discussion "NLP" | Vitaliy Radchenko, Thomas Wolf, Braden Hancock, Maksym Bevza | 40m |
Panelists report concrete moves: pretrained transformers (BERT, ELMo) drive big gains, HuggingFace’s Transformers library ships models used in benchmarks… | ||
| nothing matches | ||