Timezone: »
We introduce the unbounded depth neural network (UDN), an infinitely deep probabilistic model that adapts its complexity to the training data. The UDN contains an infinite sequence of hidden layers and places an unbounded prior on a truncation L, the layer from which it produces its data. Given a dataset of observations, the posterior UDN provides a conditional distribution of both the parameters of the infinite neural network and its truncation. We develop a novel variational inference algorithm to approximate this posterior, optimizing a distribution of the neural network weights and of the truncation depth L, and without any upper limit on L. To this end, the variational family has a special structure: it models neural network weights of arbitrary depth, and it dynamically creates or removes free variational parameters as its distribution of the truncation is optimized. (Unlike heuristic approaches to model search, it is solely through gradient-based optimization that this algorithm explores the space of truncations.) We study the UDN on real and synthetic data. We find that the UDN adapts its posterior depth to the dataset complexity; it outperforms standard neural networks of similar computational complexity; and it outperforms other approaches to infinite-depth neural networks.
Author Information
Achille Nazaret (Columbia University)
David Blei (Columbia University)
David Blei is a Professor of Statistics and Computer Science at Columbia University, and a member of the Columbia Data Science Institute. His research is in statistical machine learning, involving probabilistic topic models, Bayesian nonparametric methods, and approximate posterior inference algorithms for massive data. He works on a variety of applications, including text, images, music, social networks, user behavior, and scientific data. David has received several awards for his research, including a Sloan Fellowship (2010), Office of Naval Research Young Investigator Award (2011), Presidential Early Career Award for Scientists and Engineers (2011), Blavatnik Faculty Award (2013), and ACM-Infosys Foundation Award (2013). He is a fellow of the ACM.
Related Events (a corresponding poster, oral, or spotlight)
-
2022 Spotlight: Variational Inference for Infinitely Deep Neural Networks »
Wed. Jul 20th 06:20 -- 06:25 PM Room Room 301 - 303
More from the Same Authors
-
2022 : Optimization-based Causal Estimation from Heterogenous Environments »
Mingzhang Yin · Yixin Wang · David Blei -
2023 : Causal-structure Driven Augmentations for Text OOD Generalization »
Amir Feder · Yoav Wald · Claudia Shi · Suchi Saria · David Blei -
2023 : Practical and Asymptotically Exact Conditional Sampling in Diffusion Models »
Brian Trippe · Luhuan Wu · Christian Naesseth · David Blei · John Cunningham -
2022 : Reconstructing the Universe with Variational self-Boosted Sampling »
Chirag Modi · Yin Li · David Blei -
2022 : Probabilistic basis decomposition for characterizing temporal dynamics of gene expression »
Achille Nazaret -
2021 Poster: Unsupervised Representation Learning via Neural Activation Coding »
Yookoon Park · Sangho Lee · Gunhee Kim · David Blei -
2021 Poster: A Proxy Variable View of Shared Confounding »
Yixin Wang · David Blei -
2021 Spotlight: A Proxy Variable View of Shared Confounding »
Yixin Wang · David Blei -
2021 Oral: Unsupervised Representation Learning via Neural Activation Coding »
Yookoon Park · Sangho Lee · Gunhee Kim · David Blei -
2020 Poster: Stochastic Flows and Geometric Optimization on the Orthogonal Group »
Krzysztof Choromanski · David Cheikhi · Jared Quincy Davis · Valerii Likhosherstov · Achille Nazaret · Achraf Bahamou · Xingyou Song · Mrugank Akarte · Jack Parker-Holder · Jacob Bergquist · Yuan Gao · Aldo Pacchiano · Tamas Sarlos · Adrian Weller · Vikas Sindhwani -
2018 Poster: Noisin: Unbiased Regularization for Recurrent Neural Networks »
Adji Bousso Dieng · Rajesh Ranganath · Jaan Altosaar · David Blei -
2018 Oral: Noisin: Unbiased Regularization for Recurrent Neural Networks »
Adji Bousso Dieng · Rajesh Ranganath · Jaan Altosaar · David Blei -
2018 Poster: Augment and Reduce: Stochastic Inference for Large Categorical Distributions »
Francisco Ruiz · Michalis Titsias · Adji Bousso Dieng · David Blei -
2018 Poster: Black Box FDR »
Wesley Tansey · Yixin Wang · David Blei · Raul Rabadan -
2018 Oral: Augment and Reduce: Stochastic Inference for Large Categorical Distributions »
Francisco Ruiz · Michalis Titsias · Adji Bousso Dieng · David Blei -
2018 Oral: Black Box FDR »
Wesley Tansey · Yixin Wang · David Blei · Raul Rabadan -
2017 Workshop: Implicit Generative Models »
Rajesh Ranganath · Ian Goodfellow · Dustin Tran · David Blei · Balaji Lakshminarayanan · Shakir Mohamed -
2017 Poster: Robust Probabilistic Modeling with Bayesian Data Reweighting »
Yixin Wang · Alp Kucukelbir · David Blei -
2017 Poster: Evaluating Bayesian Models with Posterior Dispersion Indices »
Alp Kucukelbir · Yixin Wang · David Blei -
2017 Poster: Zero-Inflated Exponential Family Embeddings »
Liping Liu · David Blei -
2017 Talk: Zero-Inflated Exponential Family Embeddings »
Liping Liu · David Blei -
2017 Talk: Evaluating Bayesian Models with Posterior Dispersion Indices »
Alp Kucukelbir · Yixin Wang · David Blei -
2017 Talk: Robust Probabilistic Modeling with Bayesian Data Reweighting »
Yixin Wang · Alp Kucukelbir · David Blei