Share This Author
Deep Canonical Correlation Analysis
DCCA is introduced, a method to learn complex nonlinear transformations of two views of data such that the resulting representations are highly linearly correlated and Parameters of both transformations are jointly learned to maximize the (regularized) total correlation.
A gentle tutorial of the em algorithm and its application to parameter estimation for Gaussian mixture and hidden Markov models
- J. Bilmes
- Computer Science
The abstract form of the EM algorithm as it is often given in the literature is described and the EM parameter estimation procedure is developed for two applications: 1) finding the parameters of a mixture of Gaussian densities, and 2) finding a hidden Markov model (HMM) for both discrete and Gaussian mixture observation models.
On Deep Multi-View Representation Learning
This work finds an advantage for correlation-based representation learning, while the best results on most tasks are obtained with the new variant, deep canonically correlated autoencoders (DCCAE).
A Class of Submodular Functions for Document Summarization
A class of submodular functions meant for document summarization tasks which combine two terms, one which encourages the summary to be representative of the corpus, and the other which positively rewards diversity, which means that an efficient scalable greedy optimization scheme has a constant factor guarantee of optimality.
An integrated encyclopedia of DNA elements in the human genome
Panel C shows several SNPs associated with Crohn’s disease and other inflammatory diseases that reside in a large gene desert on chromosome 5, along with some epigenetic features suggestive of function.
Unsupervised pattern discovery in human chromatin structure through genomic segmentation
- M. M. Hoffman, Orion J. Buske, Jie Wang, Z. Weng, J. Bilmes, William Stafford Noble
- BiologyNature Methods
- 22 September 2013
Segway, a dynamic Bayesian network method, was trained simultaneously on chromatin data from multiple experiments, including positions of histone modifications, transcription-factor binding and open chromatin, all derived from a human chronic myeloid leukemia cell line to identify patterns associated with transcription start sites, gene ends, enhancers, transcriptional regulator CTCF-binding regions and repressed regions.
Integrative annotation of chromatin elements from ENCODE data
- M. M. Hoffman, J. Ernst, William Stafford Noble
- Computer Science, BiologyNucleic acids research
- 5 December 2012
These methods rediscover and summarize diverse aspects of chromatin architecture, elucidate the interplay between chromatin activity and RNA transcription, and reveal that a large proportion of the genome lies in a quiescent state, even across multiple cell types.
MVA Processing of Speech Features
It is argued and demonstrated that MVA works better when applied to the zeroth-order cepstral coefficient than to log energy, that M VA works better in the cEPstral domain, and that an ARMA filter is better than either a designed finite impulse response filter or a data-driven filter.
Multi-document Summarization via Budgeted Maximization of Submodular Functions
It is shown, both theoretically and empirically, that a modified greedy algorithm can efficiently solve the budgeted submodular maximization problem near-optimally, and derive new approximation bounds in doing so.
On Mixup Training: Improved Calibration and Predictive Uncertainty for Deep Neural Networks
- S. Thulasidasan, Gopinath Chennupati, J. Bilmes, Tanmoy Bhattacharya, S. Michalak
- Computer ScienceNeurIPS
- 27 May 2019
DNNs trained with mixup are significantly better calibrated and are less prone to over-confident predictions on out-of-distribution and random-noise data, suggesting that mixup be employed for classification tasks where predictive uncertainty is a significant concern.