AIMlogo

Adaptive Integrated Microsystems Laboratory

"Pushing the Limits of Hybrid Signal Processing"

Our research explores new frontiers in non-conventional signal processing techniques on CMOS and hybrid substrates. These include approaching limits of energy efficiency, sensing and resolution by exploiting computational primitives inherent in the physics of devices, sensors and the underlying noise processes and adaptation.

Home       Members       Research       Publications       Patents       Outreach       AIM-Software

Licensing Terms: This program is granted free of charge for research and education purposes. However you must obtain a license from the author to use it for commercial purposes. The software must not be modified and distributed without prior permission of the author. By using the AIM Lab software you agree to the licensing terms:

NO WARRANTY: BECAUSE THE PROGRAM IS LICENSED FREE OF CHARGE, THERE IS NO WARRANTY FOR THE PROGRAM, TO THE EXTENT PERMITTED BY APPLICABLE LAW. EXCEPT WHEN OTHERWISE STATED IN WRITING THE COPYRIGHT HOLDERS AND/OR OTHER PARTIES PROVIDE THE PROGRAM "AS IS" WITHOUT WARRANTY OF ANY KIND, EITHER EXPRESSED OR IMPLIED, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE. THE ENTIRE RISK AS TO THE QUALITY AND PERFORMANCE OF THE PROGRAM IS WITH YOU. SHOULD THE PROGRAM PROVE DEFECTIVE, YOU ASSUME THE COST OF ALL NECESSARY SERVICING, REPAIR OR CORRECTION. IN NO EVENT UNLESS REQUIRED BY APPLICABLE LAW OR AGREED TO IN WRITING WILL ANY COPYRIGHT HOLDER, OR ANY OTHER PARTY WHO MAY MODIFY AND/OR REDISTRIBUTE THE PROGRAM, BE LIABLE TO YOU FOR DAMAGES, INCLUDING ANY GENERAL, SPECIAL, INCIDENTAL OR CONSEQUENTIAL DAMAGES ARISING OUT OF THE USE OR INABILITY TO USE THE PROGRAM (INCLUDING BUT NOT LIMITED TO LOSS OF DATA OR DATA BEING RENDERED INACCURATE OR LOSSES SUSTAINED BY YOU OR THIRD PARTIES OR A FAILURE OF THE PROGRAM TO OPERATE WITH ANY OTHER PROGRAMS), EVEN IF SUCH HOLDER OR OTHER PARTY HAS BEEN ADVISED OF THE POSSIBILITY OF SUCH DAMAGES.

Acknowledgemtns: The open-source software development effort has been supported in part by research grants from the U.S. National Science Foundation.

Gini-Support Vector Machine

Gini-SVM is a multi-class classifier that directly produces conditional probability estimates and hence can be used for data-fusion and for integrating with other probabilistic models like Hidden Markov Models or Gaussian Mixture Models.

  • Dedicated Gini-SVM website for updates and documentation can be found here.
  • The theory and learning algorithm underlying Gini-SVM can be found in the following JMLR paper.
  • MATLAB based Gini-SVM toolkit (Version 1.1) is available (here). The toolkit implements a randomized but slower version of the Gini-SVM algorithm suitable for smaller data-sets.
  • A C++ version of the Gini-SVM toolkit (Version 1.1) is available (here). The toolkit implements kernel caching and speed-up heuristics and is suitable for larger data-sets.

SPARK Robust Speech Feature Extractor

SParse Auditory Reproducing Kernels based speech features are hierarchical features inspired from human auditory processing systems. The method is as computationally efficient as the mel-frequency cepstral coefficient features but deliver significant improvement in recognition performance (Speech and speaker) under noisy, real-life recording conditions.

  • Open-source code (MATLAB version) for SPARK (version 1.1) features can be downloaded (here). Please see test_SPARK.m for an example on how to run the code.
  • Theoretical and experimental results using the SPARK features can be found in A. Fazel, S. Chakrabartty, ``Sparse Auditory Reproducing Kernels (SPARK) Features for Noise-Robust Speech Recognition, IEEE Transactions on Audio, Speech, and Language Processing, vol. 20, no:4, 2012. Pre-print of the paper is available (here).

FAST: Factor-graph Analysis of Stochastic CircuiTs.

FAST is a Monte-Carlo analysis tools for estimating the reliability of circuits whose components are stochastic elements, for example protiens and biomolecules.

  • The documentaion describing the FAST interface is provided in the following (paper).
  • Paper describing the biomolecular analysis technique using factor-graphs can be downloaded (here).
  • Open-source code for FAST (version 1.1) can be downloaded (here).

Margin Propagation (MP) Synthesis and Simulation

Margin propagation is a technique for synthesizing piece-wise linear (PWL) analog computational circuits that are bias-scalable and rely only on universal principles like the Kirchoff's current law (KCL) or rectification or thresholding. Thus, unlike translinear circuits MP circuits can functionally similar in weak, moderate and strong inversion.

  • Paper describing the MP-based synthesis can be downloaded (here).
  • Paper describing MP-based LDPC decoder can be downloaded (here).
  • Open-source code for MarginLDPC (version 1.0) can be downloaded (here).

Power-point Circuit Drawing Tool

This is a power-point slide which is a compilation of different circuit elements (resistors, capacitors,...,MOS transistors) that can be used to produce publication quality circuits drawings.

  • Power-point file can be downloaded (here).

Questions and comments: shantanu at msu dot edu