Explainable AI via Learning to Optimize (2204.14174v2)
Abstract: Indecipherable black boxes are common in ML, but applications increasingly require explainable artificial intelligence (XAI). The core of XAI is to establish transparent and interpretable data-driven algorithms. This work provides concrete tools for XAI in situations where prior knowledge must be encoded and untrustworthy inferences flagged. We use the "learn to optimize" (L2O) methodology wherein each inference solves a data-driven optimization problem. Our L2O models are straightforward to implement, directly encode prior knowledge, and yield theoretical guarantees (e.g. satisfaction of constraints). We also propose use of interpretable certificates to verify whether model inferences are trustworthy. Numerical examples are provided in the applications of dictionary-based signal recovery, CT imaging, and arbitrage trading of cryptoassets. Code and additional documentation can be found at https://xai-l2o.research.typal.academy.
- “Peeking inside the black-box: a survey on explainable artificial intelligence (XAI)” In IEEE access 6 IEEE, 2018, pp. 52138–52160
- Jonas Adler, Holger Kohr and Ozan Öktem “Operator Discretization Library (ODL)”, 2017
- Brandon Amos “Tutorial on amortized optimization for learning to optimize over continuous domains” In arXiv preprint arXiv:2202.00665, 2022
- Mariano Anaya “Clean Code in Python: Refactor your legacy code base” Packt Publishing Ltd, 2018
- “Improved price oracles: Constant function market makers” In Proceedings of the 2nd ACM Conference on Advances in Financial Technologies, 2020, pp. 80–91
- “Constant function market makers: Multi-asset trades via convex optimization” In arXiv preprint arXiv:2107.12484, 2021
- “Optimal Routing for Constant Function Market Makers”, 2021
- “FactSheets: Increasing trust in AI services through supplier’s declarations of conformity” In IBM Journal of Research and Development 63.4/5 IBM, 2019, pp. 6–1
- “Explainable Artificial Intelligence (XAI): Concepts, taxonomies, opportunities and challenges toward responsible AI” In Information Fusion 58 Elsevier, 2020, pp. 82–115
- “On pixel-wise explanations for non-linear classifier decisions by layer-wise relevance propagation” In PloS one 10.7 Public Library of Science San Francisco, CA USA, 2015, pp. e0130140
- Shaojie Bai, J Zico Kolter and Vladlen Koltun “Deep equilibrium models” In arXiv preprint arXiv:1909.01377, 2019
- Shaojie Bai, Vladlen Koltun and J Zico Kolter “Multiscale deep equilibrium models” In arXiv preprint arXiv:2006.08656, 2020
- Shaojie Bai, Vladlen Koltun and Zico Kolter “Stabilizing Equilibrium Models by Jacobian Regularization” In Proceedings of the 38th International Conference on Machine Learning 139, Proceedings of Machine Learning Research PMLR, 2021, pp. 554–565 URL: https://proceedings.mlr.press/v139/bai21b.html
- Amir Beck “First-Order Methods in Optimization” SIAM, 2017
- Emmanuel J Candès, Justin Romberg and Terence Tao “Robust uncertainty principles: Exact signal reconstruction from highly incomplete frequency information” In IEEE Transactions on information theory 52.2 IEEE, 2006, pp. 489–509
- “On the local behavior of spaces of natural images” In International journal of computer vision 76.1 Springer, 2008, pp. 1–12
- Tony Chan, Antonio Marquina and Pep Mulet “High-order total variation-based image restoration” In SIAM Journal on Scientific Computing 22.2 SIAM, 2000, pp. 503–516
- “Learning to optimize: A primer and a benchmark” In arXiv preprint arXiv:2103.12828, 2021
- “Theoretical linear convergence of unfolded ISTA and its practical weights and thresholds” In arXiv preprint arXiv:1808.10038, 2018
- Patrick L Combettes and Jean-Christophe Pesquet “Lipschitz certificates for layered network structures driven by averaged activation operators” In SIAM Journal on Mathematics of Data Science 2.2 SIAM, 2020, pp. 529–557
- “Flash boys 2.0: Frontrunning, transaction reordering, and consensus instability in decentralized exchanges” In arXiv preprint arXiv:1904.05234, 2019
- “A three-operator splitting scheme and its optimization applications” In Set-valued and variational analysis 25.4 Springer, 2017, pp. 829–858
- “On the global and linear convergence of the generalized alternating direction method of multipliers” In Journal of Scientific Computing 66.3 Springer, 2016, pp. 889–916
- Filip Karlo Došilović, Mario Brčić and Nikica Hlupić “Explainable artificial intelligence: A survey” In 2018 41st International convention on information and communication technology, electronics and microelectronics (MIPRO), 2018, pp. 0210–0215 IEEE
- “Variable selection via nonconcave penalized likelihood and its oracle properties” In Journal of the American statistical Association 96.456 Taylor & Francis, 2001, pp. 1348–1360
- “JFB: Jacobian-free backpropagation for implicit networks” In arXiv preprint arXiv:2103.12803, 2021
- “On the properties of the softmax function with application in game theory and reinforcement learning” In arXiv preprint arXiv:1704.00805, 2017
- “Compressive sensing for missing data imputation in noise robust speech recognition” In IEEE Journal of selected topics in Signal Processing 4.2 IEEE, 2010, pp. 272–287
- “On Training Implicit Models” In Thirty-Fifth Conference on Neural Information Processing Systems, 2021
- Davis Gilton, Gregory Ongie and Rebecca Willett “Deep equilibrium architectures for inverse problems in imaging” In arXiv preprint arXiv:2102.07944, 2021
- “Learning fast approximations of sparse coding” In Proceedings of the 27th international conference on international conference on machine learning, 2010, pp. 399–406
- “Feasibility-based fixed point networks” In arXiv preprint arXiv:2104.14090, 2021
- “Learn to Predict Equilibria via Fixed Point Networks” In arXiv preprint arXiv:2106.00906, 2021
- Eyal Hertzog, Guy Benartzi and Galia Benartzi “Bancor protocol” storage.googleapis. com/website-bancor/2018/04/01ba8253-bancor_protocol_whitepaper_en.pdf (accessed on April 24 2022) In White Paper, 2017
- Zhichun Huang, Shaojie Bai and J Zico Kolter “Implicit2: Implicit Layers for Implicit Representations” In Advances in Neural Information Processing Systems 34, 2021
- “Super-resolution ct image reconstruction based on dictionary learning and sparse representation” In Scientific reports 8.1 Nature Publishing Group, 2018, pp. 1–10
- “Deep convolutional neural network for inverse problems in imaging” In IEEE Transactions on Image Processing 26.9 IEEE, 2017, pp. 4509–4522
- Sep Kamvar, Marek Olszewski and Rene Reinsberg “Celo: A multi-asset cryptographic protocol for decentralized social payments” storage. googleapis. com/celo whitepapers/Celo A Multi Asset Cryptographic Protocol for Decentralized Social Payments.pdf In White Paper, 2019
- Diederik P Kingma and Jimmy Ba “Adam: A method for stochastic optimization” In arXiv preprint arXiv:1412.6980, 2014
- “Music Source Separation with Deep Equilibrium Models” In arXiv preprint arXiv:2110.06494, 2021
- M.A. Krasnosel’skii “Two remarks about the method of successive approximations” In Uspekhi Mat. Nauk 10, 1955, pp. 123–127
- Ann B Lee, Kim S Pedersen and David Mumford “The nonlinear statistics of high-contrast patches in natural images” In International Journal of Computer Vision 54.1-3 Springer, 2003, pp. 83–103
- “The LoDoPaB-CT dataset: A benchmark dataset for low-dose CT reconstruction methods” In arXiv preprint arXiv:1910.01113, 2019
- “Alternating the population and control neural networks to solve high-dimensional stochastic mean-field games” In Proceedings of the National Academy of Sciences 118.31 National Acad Sciences, 2021
- “ALISTA: Analytic weights are as good as learned weights in LISTA” In International Conference on Learning Representations (ICLR), 2019
- “Trading and arbitrage in cryptocurrency markets” In Journal of Financial Economics 135.2 Elsevier, 2020, pp. 293–319
- “Model cards for model reporting” In Proceedings of the conference on fairness, accountability, and transparency, 2019, pp. 220–229
- Vishal Monga, Yuelong Li and Yonina C Eldar “Algorithm unrolling: Interpretable, efficient deep learning for signal and image processing” In IEEE Signal Processing Magazine 38.2 IEEE, 2021, pp. 18–44
- “Layer-wise relevance propagation: an overview” In Explainable AI: interpreting, explaining and visualizing deep learning Springer, 2019, pp. 193–209
- “The Care Label Concept: A Certification Suite for Trustworthy and Resource-Aware Machine Learning” In arXiv preprint arXiv:2106.00512, 2021
- “Yes We Care!–Certification for Machine Learning Methods through the Care Label Framework” In arXiv preprint arXiv:2105.10197, 2021
- Stanley Osher, Zuoqiang Shi and Wei Zhu “Low dimensional manifold model for image processing” In SIAM Journal on Imaging Sciences 10.4 SIAM, 2017, pp. 1669–1690
- “Pytorch: An imperative style, high-performance deep learning library” In Advances in neural information processing systems 32, 2019, pp. 8026–8037
- Gabriel Peyré “Image processing with nonlocal spectral bases” In Multiscale Modeling & Simulation 7.2 SIAM, 2008, pp. 703–730
- Gabriel Peyré “Manifold models for signals and images” In Computer vision and image understanding 113.2 Elsevier, 2009, pp. 249–260
- MakerDAO Project “The Maker Protocol: MakerDAO’s Multi-Collateral Dai (MCD) System” storage. googleapis. com/celo whitepapers/Celo A Multi Asset Cryptographic Protocol for Decentralized Social Payments.pdf In White Paper, 2020
- Maziar Raissi, Paris Perdikaris and George E Karniadakis “Physics-informed neural networks: A deep learning framework for solving forward and inverse problems involving nonlinear partial differential equations” In Journal of Computational Physics 378 Elsevier, 2019, pp. 686–707
- Wolfgang Ring “Structural properties of solutions to total variation regularization problems” In ESAIM: Mathematical Modelling and Numerical Analysis 34.4 EDP Sciences, 2000, pp. 799–810
- “A machine learning framework for solving high-dimensional mean field game and mean field control problems” In Proceedings of the National Academy of Sciences 117.17 National Acad Sciences, 2020, pp. 9183–9193
- “Large-Scale Convex Optimization: Algorithm Designs via Monotone Operators” Cambridge University Press, 2022 URL: %5Curl%7Bhttps://large-scale-book.mathopt.com%7D
- “Towards explainable artificial intelligence” In Explainable AI: interpreting, explaining and visualizing deep learning Springer, 2019, pp. 5–22
- Fabian Schär “Decentralized finance: On blockchain-and smart contract-based financial markets” In FRB of St. Louis Review, 2021
- “Model-based deep learning” In arXiv preprint arXiv:2012.08405, 2020
- KV Siddamal, Shobha P Bhat and VS Saroja “A survey on compressive sensing” In 2015 2nd International Conference on Electronics and Communication Systems (ICECS), 2015, pp. 639–643 IEEE
- Michael Van Lent, William Fisher and Michael Mancuso “An explainable artificial intelligence system for small-unit tactical behavior” In Proceedings of the national conference on artificial intelligence, 2004, pp. 900–907 Menlo Park, CA; Cambridge, MA; London; AAAI Press; MIT Press; 1999
- “0x: An open protocol for decentralized exchange on the Ethereum blockchain” github.com/0xProject/whitepaper In White Paper, 2017
- “Sok: Decentralized finance (defi)” In arXiv preprint arXiv:2101.08778, 2021
- “Low-dose X-ray CT reconstruction via dictionary learning” In IEEE transactions on medical imaging 31.9 IEEE, 2012, pp. 1682–1697
- Yi Zhang, Xiaohong Chen and Daejun Park “Formal specification of constant product (xy= k) market maker model and implementation” In White Paper, 2018
- “A survey of sparse representation: algorithms and applications” In IEEE access 3 IEEE, 2015, pp. 490–530
Paper Prompts
Sign up for free to create and run prompts on this paper using GPT-5.
Top Community Prompts
Collections
Sign up for free to add this paper to one or more collections.