no code implementations • 24 Oct 2023 • Benyamin Haghi, Lin Ma, Sahin Lale, Anima Anandkumar, Azita Emami
We present an integrated approach by combining analog computing and deep learning for electrocardiogram (ECG) arrhythmia classification.
no code implementations • 19 Jan 2023 • Peter I Renn, Cong Wang, Sahin Lale, Zongyi Li, Anima Anandkumar, Morteza Gharib
The learned FNO solution operator can be evaluated in milliseconds, potentially enabling faster-than-real-time modeling for predictive flow control in physical systems.
no code implementations • 17 Jun 2022 • Taylan Kargin, Sahin Lale, Kamyar Azizzadenesheli, Anima Anandkumar, Babak Hassibi
By carefully prescribing an early exploration strategy and a policy update rule, we show that TS achieves order-optimal regret in adaptive control of multidimensional stabilizable LQRs.
no code implementations • 3 Jun 2022 • Sahin Lale, Yuanyuan Shi, Guannan Qu, Kamyar Azizzadenesheli, Adam Wierman, Anima Anandkumar
However, current reinforcement learning (RL) methods lack stabilization guarantees, which limits their applicability for the control of safety-critical systems.
no code implementations • 3 Jun 2022 • Oron Sabag, Sahin Lale, Babak Hassibi
The key techniques that underpin our explicit solution is a reduction of the control problem to a Nehari problem, along with a novel factorization of the clairvoyant controller's cost.
no code implementations • 22 Feb 2022 • Navid Azizan, Sahin Lale, Babak Hassibi
RMD starts with a standard cost which is the sum of the training loss and a convex regularizer of the weights.
1 code implementation • 14 Dec 2021 • Kevin Huang, Sahin Lale, Ugo Rosolia, Yuanyuan Shi, Anima Anandkumar
It then uses the top trajectories as initialization for gradient descent and applies gradient updates to each of these trajectories to find the optimal action sequence.
no code implementations • 26 Aug 2021 • Sahin Lale, Kamyar Azizzadenesheli, Babak Hassibi, Anima Anandkumar
Using these guarantees, we design adaptive control algorithms for unknown ARX systems with arbitrary strongly convex or convex quadratic regulating costs.
no code implementations • 4 May 2021 • Oron Sabag, Gautam Goel, Sahin Lale, Babak Hassibi
Motivated by competitive analysis in online learning, as a criterion for controller design we introduce the dynamic regret, defined as the difference between the LQR cost of a causal controller (that has only access to past disturbances) and the LQR cost of the \emph{unique} clairvoyant one (that has also access to future disturbances) that is known to dominate all other controllers.
no code implementations • 29 Apr 2021 • Guannan Qu, Yuanyuan Shi, Sahin Lale, Anima Anandkumar, Adam Wierman
In this work, we propose an efficient online control algorithm, COvariance Constrained Online Linear Quadratic (COCO-LQ) control, that guarantees input-to-state stability for a large class of LTV systems while also minimizing the control cost.
no code implementations • 8 Dec 2020 • Sahin Lale, Oguzhan Teke, Babak Hassibi, Anima Anandkumar
In this model, each state variable is updated randomly and asynchronously with some probability according to the underlying system dynamics.
no code implementations • 23 Jul 2020 • Sahin Lale, Kamyar Azizzadenesheli, Babak Hassibi, Anima Anandkumar
In this work, we study model-based reinforcement learning (RL) in unknown stabilizable linear dynamical systems.
Model-based Reinforcement Learning reinforcement-learning +1
no code implementations • NeurIPS 2020 • Sahin Lale, Kamyar Azizzadenesheli, Babak Hassibi, Anima Anandkumar
We study the problem of system identification and adaptive control in partially observable linear dynamical systems.
no code implementations • 12 Mar 2020 • Sahin Lale, Kamyar Azizzadenesheli, Babak Hassibi, Anima Anandkumar
We study the problem of adaptive control in partially observable linear quadratic Gaussian control systems, where the model dynamics are unknown a priori.
no code implementations • 31 Jan 2020 • Sahin Lale, Kamyar Azizzadenesheli, Babak Hassibi, Anima Anandkumar
We propose a novel way to decompose the regret and provide an end-to-end sublinear regret upper bound for partially observable linear quadratic control.
no code implementations • 25 Sep 2019 • Navid Azizan, Sahin Lale, Babak Hassibi
On the theory side, we show that in the overparameterized nonlinear setting, if the initialization is close enough to the manifold of global optima, SMD with sufficiently small step size converges to a global minimum that is approximately the closest global minimum in Bregman divergence, thus attaining approximate implicit regularization.
1 code implementation • 10 Jun 2019 • Navid Azizan, Sahin Lale, Babak Hassibi
Most modern learning problems are highly overparameterized, meaning that there are many more parameters than the number of training data points, and as a result, the training loss may have infinitely many global minima (parameter vectors that perfectly interpolate the training data).
no code implementations • 28 Jan 2019 • Sahin Lale, Kamyar Azizzadenesheli, Anima Anandkumar, Babak Hassibi
We modify the image classification task into the SLB setting and empirically show that, when a pre-trained DNN provides the high dimensional feature representations, deploying PSLB results in significant reduction of regret and faster convergence to an accurate model compared to state-of-art algorithm.