Abstract
In this paper we aim at facilitating generalization for deep networks while supporting interpretability of the learned representations. Towards this goal, we propose a clustering based regularization that encourages parsimonious representations. Our k-means style objective is easy to optimize and flexible, supporting various forms of clustering, such as sample clustering, spatial clustering, as well as co-clustering. We demonstrate the effectiveness of our approach on the tasks of unsupervised learning, classification, fine grained categorization, and zero-shot learning.
| Original language | English (US) |
|---|---|
| Pages (from-to) | 5092-5100 |
| Number of pages | 9 |
| Journal | Advances in Neural Information Processing Systems |
| State | Published - Jan 1 2016 |
| Event | 30th Annual Conference on Neural Information Processing Systems, NIPS 2016 - Barcelona, Spain Duration: Dec 5 2016 → Dec 10 2016 |
Fingerprint
ASJC Scopus subject areas
- Computer Networks and Communications
- Information Systems
- Signal Processing
Cite this
Learning deep parsimonious representations. / Liao, Renjie; Schwing, Alexander Gerhard; Zemel, Richard S.; Urtasun, Raquel.
In: Advances in Neural Information Processing Systems, 01.01.2016, p. 5092-5100.Research output: Contribution to journal › Conference article
}
TY - JOUR
T1 - Learning deep parsimonious representations
AU - Liao, Renjie
AU - Schwing, Alexander Gerhard
AU - Zemel, Richard S.
AU - Urtasun, Raquel
PY - 2016/1/1
Y1 - 2016/1/1
N2 - In this paper we aim at facilitating generalization for deep networks while supporting interpretability of the learned representations. Towards this goal, we propose a clustering based regularization that encourages parsimonious representations. Our k-means style objective is easy to optimize and flexible, supporting various forms of clustering, such as sample clustering, spatial clustering, as well as co-clustering. We demonstrate the effectiveness of our approach on the tasks of unsupervised learning, classification, fine grained categorization, and zero-shot learning.
AB - In this paper we aim at facilitating generalization for deep networks while supporting interpretability of the learned representations. Towards this goal, we propose a clustering based regularization that encourages parsimonious representations. Our k-means style objective is easy to optimize and flexible, supporting various forms of clustering, such as sample clustering, spatial clustering, as well as co-clustering. We demonstrate the effectiveness of our approach on the tasks of unsupervised learning, classification, fine grained categorization, and zero-shot learning.
UR - http://www.scopus.com/inward/record.url?scp=85018904825&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=85018904825&partnerID=8YFLogxK
M3 - Conference article
AN - SCOPUS:85018904825
SP - 5092
EP - 5100
JO - Advances in Neural Information Processing Systems
JF - Advances in Neural Information Processing Systems
SN - 1049-5258
ER -