Deep Learning Indaba 2018
2018-09-09..2018-09-14
Stellenbosch, South Africa
| title | author | length |
|---|---|---|
| Evaluating Model Explainability Methods | Sara Hooker | 01h 05m |
Paper evaluates pixel-level feature importance methods (gradients, integrated gradients, guided backprop) by forcing a controlled mean-shift invariance test… | ||
| Learning Representations for Visual Search | Naila Murray | 42m |
Describes converting eye fixations into a probability distribution of fixations by Gaussian‑blurring fixation points, trains models with distributional losses… | ||
| Acoustic Word Embeddings for Low Resource Speech Processing | Herman Kamper | 01h 02m |
Proposes that variable-duration speech segments map to fixed vectors so repeated words can be found without transcripts, using acoustic word embeddings trained… | ||
| AI Ethics, Strategic Decisioning and Game Theory | Osonde Osoba | 47m |
Argues that deploying algorithms is a social experiment and evaluates them with Belmont principles made concrete for practice, showing how facial‑recognition… | ||
| Advanced Reinforcement Learning & Data Science for Social Impact | Vukosi Marivate | 47m |
Covers concrete methods and prototypes: uses inverse reinforcement learning and multi‑agent ensembles, builds a probabilistic model‑based RL pipeline to… | ||
| nothing matches | ||