SPECIAL NOTICE
99 -- TECHNOLOGY/BUSINESS OPPORTUNITY MimicGAN: An algorithmic protocol which enables self-correction against unknown corruptions and artifacts in datasets not seen during training
- Notice Date
- 4/2/2020 8:16:39 AM
- Notice Type
- Special Notice
- Contracting Office
- LLNS � DOE CONTRACTOR Livermore CA 94551 USA
- ZIP Code
- 94551
- Solicitation Number
- FBO456-20
- Response Due
- 5/14/2020 9:00:00 PM
- Archive Date
- 05/16/2020
- Point of Contact
- Connie Pitcock, Phone: 925-422-1072
- E-Mail Address
-
pitcock11@llnl.gov
(pitcock11@llnl.gov)
- Description
- Opportunity: Lawrence Livermore National Laboratory (LLNL), operated by the Lawrence Livermore National Security (LLNS), LLC under contract no. DE-AC52-07NA27344 (Contract 44) with the U.S. Department of Energy (DOE), is offering the opportunity to license its MimicGAN software tool for commercialization and use. Background: The algorithmic and computational advances over the past decade have yielded significant performance boosts of machine learning (ML) models. As a result, we are witnessing rapid adoption of ML algorithms like classifiers and detectors, in real-world applications. Several industries use ML inside their workflows and manufacturing processes to improve on metrics like speed, scale, and productivity. However, ML models are not �battle-ready� in the sense that they can be easily broken. The fundamental assumption that is made in deploying virtually any current ML model is that the training data covers all possible variations in the field (test) � but this is rarely the case. Even if care is taken to ensure they are similar, ambient noise (like lighting changes, uncontrollable occlusions, etc.) from the sensing environment or faulty sensors can cause even the most sophisticated algorithms to fail. There is also evidence that ML algorithms are vulnerable to carefully designed adversarial corruptions which are invisible to the human observer but can result in incorrect results. The challenge in these scenarios is often that we do not have knowledge of the kind of corruption in advance making it hard to detect and correct for it. Description:� MimicGAN represents a new generation of methods that can �self-correct� for unseen corruptions in the data out in the field. This is particularly useful for systems that need to be deployed autonomously without needing constant intervention such as Automated Driver Assistance Systems. MimicGAN achieves this by treating every test sample as �corrupt� by default. The goal is to determine (a) the clean image and (b) the corruption both of which are unknown to the system at test time. MimicGAN solves this by making alternating guesses between what the clean sample should look like and what corruption might make it look like the observed corrupted sample. If there is no corruption at all, MimicGAN simply learns the corruption to be an identity transform � i.e., no corruption. By operating in real-time with as few as 10 corrupted samples, MimicGAN serves as a very powerful pre-processing step to ML models deployed in the wild. It can even correct for adversarial corruptions better than most existing methods. Preliminary results on images have demonstrated that a wide variety of significant unknown corruptions like missing pixels or geometric transformations (rotations, scale changes, noise, pixel intensity changes, etc.) can be accurately estimated from a small handful of examples. Furthermore, MimicGAN applies not only to images but to any data distribution that can be learned with modern neural networks including multi-modal scalars, spectra, time series, etc. This make MimicGAN a very flexible tool that can improve a wide variety of tasks. For example, using MimicGAN one can train models on simulated data but apply them to experiments, or use cheap low-fidelity data to inform high-fidelity models. Similarly, one may want to train from data of one healthcare system but apply the models to another, fit housing prices, energy consumption, or traffic in one city but predict in another etc. Ultimately, any application in which the test data may not be similar to the training data yet there is too little data or time to re-train a model will benefit from incorporating MimicGAN. Advantages:� MimicGAN is a Machine Learning software solution that can: Be applied to wide class of problems where one has access to a large amount of synthetic or lab data but need predictions and inferences on relatively expensive �field data� with the highest accuracy possible. Generalize ML models to data not seen during training, especially corruptions from ambient noise, faulty sensors, uncontrollable environmental artifacts. Outperform existing baselines in blind image recovery Defend against adversarial attacks � (un)intentionally introduced corruptions Potential Applications: MimicGAN is application agnostic and can be designed to work with any system that makes decisions based on sparse data. Examples in which this software is ideally targeted for are: Automated Driver Assistance Systems Health care Remote sensing Medical imaging Non-destructive evaluation Development Status: MimicGAN is fully developed and functional. Preliminary results on images have demonstrated that a wide variety of significant corruptions, i.e. missing pixels, geometric transformations, noise, etc. can be accurately estimated from a small handful of high-fidelity examples. A publication on the results can be found here: Anirudh. R., Thiagarajan, J. J., Kailkhura, B., Bremer, P.T., MimicGAN: Robust Projection onto Image Manifolds with Corruption Mimicking. �(to appear in) Int J Comput Vis�(2020). DOI:10.1007/s11263-020-01310-5 Preprint is available here: https://arxiv.org/abs/1912.07748 LLNL has filed a patent application for this algorithmic protocol. LLNL is seeking industry partners with a demonstrated ability to bring such inventions to the market. Moving critical technology beyond the Laboratory to the commercial world helps our licensees gain a competitive edge in the marketplace. All licensing activities are conducted under policies relating to the strict nondisclosure of company proprietary information.� Please visit the IPO website at https://ipo.llnl.gov/resources for more information on working with LLNL and the industrial partnering and technology transfer process. Note:� THIS IS NOT A PROCUREMENT.� Companies interested in commercializing LLNL's algorithmic protocol, MimicGAN should provide a written statement of interest, which includes the following 1.�� Company Name and address. 2.�� The name, address, and telephone number of a point of contact. 3. A description of corporate expertise and facilities relevant to commercializing this technology. Written responses should be directed to: Lawrence Livermore National Laboratory Industrial Partnerships Office P.O. Box 808, L-795 Livermore, CA� 94551-0808 Attention:� FBO 456-20 Please provide your written statement within forty-five (45) days from the date this announcement is published to ensure consideration of your interest in LLNL's algorithmic protocol, MimicGAN.
- Web Link
-
SAM.gov Permalink
(https://beta.sam.gov/opp/fda3b242c4114bc788b4dfde815b8f68/view)
- Record
- SN05608289-F 20200404/200402230207 (samdaily.us)
- Source
-
SAM.gov Link to This Notice
(may not be valid after Archive Date)
| FSG Index | This Issue's Index | Today's SAM Daily Index Page |