Nonlinear model predictive control (NMPC) is typically restricted to short, finite horizons to limit the computational burden of online optimization. As a result, global planning frameworks are frequently necessary to avoid local minima when using NMPC for navigation in complex environments. By contrast, reinforcement learning (RL) can generate policies that minimize the expected cost over an infinite-horizon and can often avoid local minima, even when operating only on current sensor measurements. However, these learned policies are usually unable to provide performance guarantees (e.g., on collision avoidance), especially when outside of the training distribution. In this paper, we augment Probably Approximately Correct NMPC (PAC-NMPC), a sampling-based stochastic NMPC algorithm capable of providing statistical guarantees of performance and safety, with an approximate perception-dependent value function trained via RL. We demonstrate in simulation that our algorithm can improve the long-term behavior of PAC-NMPC while outperforming other approaches with regards to safety for both planar car dynamics and more complex, high-dimensional fixed-wing aerial vehicle dynamics. We also demonstrate that, even when our value function is trained in simulation, our algorithm can successfully achieve statistically safe navigation on hardware using a 1/10th scale rally car in cluttered real-world environments using only current sensor information.
相關內容
Optimizing risk-averse objectives in discounted MDPs is challenging because most models do not admit direct dynamic programming equations and require complex history-dependent policies. In this paper, we show that the risk-averse {\em total reward criterion}, under the Entropic Risk Measure (ERM) and Entropic Value at Risk (EVaR) risk measures, can be optimized by a stationary policy, making it simple to analyze, interpret, and deploy. We propose exponential value iteration, policy iteration, and linear programming to compute optimal policies. Compared with prior work, our results only require the relatively mild condition of transient MDPs and allow for {\em both} positive and negative rewards. Our results indicate that the total reward criterion may be preferable to the discounted criterion in a broad range of risk-averse reinforcement learning domains.
Weighted Timed Games (WTG for short) are the most widely used model to describe controller synthesis problems involving real-time issues. Unfortunately, they are notoriously difficult, and undecidable in general. As a consequence, one-clock WTGs have attracted a lot of attention, especially because they are known to be decidable when only non-negative weights are allowed. However, when arbitrary weights are considered, despite several recent works, their decidability status was still unknown. In this paper, we solve this problem positively and show that the value function can be computed in exponential time (if weights are encoded in unary).
Large language models (LLMs) have transformed the way computers understand and process human language, but using them effectively across different organizations remains still difficult. When organizations work together to improve LLMs, they face several main challenges. First, organizations hesitate to share their valuable data with others. Second, competition between organizations creates trust problems during collaboration. Third, new privacy laws require organizations to be able to delete specific data when requested, which is especially difficult when multiple organizations are learning from shared data. Traditional federated learning approaches do not address these interconnected challenges, particularly in scenarios where participants cannot fully trust each other or the central aggregator. To overcome these limitations, we propose a hybrid blockchain-based federated learning framework that uniquely combines public and private blockchain architectures with multi-agent reinforcement learning. Our framework enables transparent sharing of model update through the public blockchain while protecting sensitive computations in private chains. Each organization operates as an intelligent agent, using Q-learning to optimize its participation strategy and resource allocation, thus aligning individual incentives with collective goals. Notably, we introduce an efficient unlearning mechanism based on Low-Rank Adaptation (LoRA) that enables selective removal of specific data contributions without compromising the model's overall performance. Through extensive experimentation on real-world datasets, we demonstrate that our framework effectively balances privacy protection, trust establishment, and regulatory compliance while maintaining high model performance.
Confidence calibration of classification models is a technique to estimate the true posterior probability of the predicted class, which is critical for ensuring reliable decision-making in practical applications. Existing confidence calibration methods mostly use statistical techniques to estimate the calibration curve from data or fit a user-defined calibration function, but often overlook fully mining and utilizing the prior distribution behind the calibration curve. However, a well-informed prior distribution can provide valuable insights beyond the empirical data under the limited data or low-density regions of confidence scores. To fill this gap, this paper proposes a new method that integrates the prior distribution behind the calibration curve with empirical data to estimate a continuous calibration curve, which is realized by modeling the sampling process of calibration data as a binomial process and maximizing the likelihood function of the binomial process. We prove that the calibration curve estimating method is Lipschitz continuous with respect to data distribution and requires a sample size of $3/B$ of that required for histogram binning, where $B$ represents the number of bins. Also, a new calibration metric ($TCE_{bpm}$), which leverages the estimated calibration curve to estimate the true calibration error (TCE), is designed. $TCE_{bpm}$ is proven to be a consistent calibration measure. Furthermore, realistic calibration datasets can be generated by the binomial process modeling from a preset true calibration curve and confidence score distribution, which can serve as a benchmark to measure and compare the discrepancy between existing calibration metrics and the true calibration error. The effectiveness of our calibration method and metric are verified in real-world and simulated data.
Advance in medical imaging is an important part in deep learning research. One of the goals of computer vision is development of a holistic, comprehensive model which can identify tumors from histology slides obtained via biopsies. A major problem that stands in the way is lack of data for a few cancer-types. In this paper, we ascertain that data augmentation using GANs can be a viable solution to reduce the unevenness in the distribution of different cancer types in our dataset. Our demonstration showed that a dataset augmented to a 50% increase causes an increase in tumor detection from 80% to 87.5%
The accurate prediction of smooth steering inputs is crucial for automotive applications because control actions with jitter might cause the vehicle system to become unstable. To address this problem in automobile lane-keeping control without the use of additional smoothing algorithms, we developed a novel soft-constrained iterative linear quadratic regulator (soft-CILQR) algorithm by integrating CILQR algorithm and a model predictive control (MPC) constraint relaxation method. We incorporated slack variables into the state and control barrier functions of the soft-CILQR solver to soften the constraints in the optimization process such that control input stabilization can be achieved in a computationally simple manner. Two types of automotive lane-keeping experiments (numerical simulations and experiments involving challenging vision-based maneuvers) were conducted with a linear system dynamics model to test the performance of the proposed soft-CILQR algorithm, and its performance was compared with that of the CILQR algorithm. In the numerical simulations, the soft-CILQR and CILQR solvers managed to drive the system toward the reference state asymptotically; however, the soft-CILQR solver obtained smooth steering input trajectories more easily than did the CILQR solver under conditions involving additive disturbances. The results of the vision-based experiments in which an ego vehicle drove in perturbed TORCS environments with various road friction settings were consistent with those of the numerical tests. The proposed soft-CILQR algorithm achieved an average runtime of 2.55 ms and is thus applicable for real-time autonomous driving scenarios.
Recurrent neural networks (RNNs) are valued for their computational efficiency and reduced memory requirements on tasks involving long sequence lengths but require high memory-processor bandwidth to train. Checkpointing techniques can reduce the memory requirements by only storing a subset of intermediate states, the checkpoints, but are still rarely used due to the computational overhead of the additional recomputation phase. This work addresses these challenges by introducing memory-efficient gradient checkpointing strategies tailored for the general class of sparse RNNs and Spiking Neural Networks (SNNs). SNNs are energy efficient alternatives to RNNs thanks to their local, event-driven operation and potential neuromorphic implementation. We use the Intelligence Processing Unit (IPU) as an exemplary platform for architectures with distributed local memory. We exploit its suitability for sparse and irregular workloads to scale SNN training on long sequence lengths. We find that Double Checkpointing emerges as the most effective method, optimizing the use of local memory resources while minimizing recomputation overhead. This approach reduces dependency on slower large-scale memory access, enabling training on sequences over 10 times longer or 4 times larger networks than previously feasible, with only marginal time overhead. The presented techniques demonstrate significant potential to enhance scalability and efficiency in training sparse and recurrent networks across diverse hardware platforms, and highlights the benefits of sparse activations for scalable recurrent neural network training.
The ability of a robot to plan complex behaviors with real-time computation, rather than adhering to predesigned or offline-learned routines, alleviates the need for specialized algorithms or training for each problem instance. Monte Carlo Tree Search is a powerful planning algorithm that strategically explores simulated future possibilities, but it requires a discrete problem representation that is irreconcilable with the continuous dynamics of the physical world. We present Spectral Expansion Tree Search (SETS), a real-time, tree-based planner that uses the spectrum of the locally linearized system to construct a low-complexity and approximately equivalent discrete representation of the continuous world. We prove SETS converges to a bound of the globally optimal solution for continuous, deterministic and differentiable Markov Decision Processes, a broad class of problems that includes underactuated nonlinear dynamics, non-convex reward functions, and unstructured environments. We experimentally validate SETS on drone, spacecraft, and ground vehicle robots and one numerical experiment, each of which is not directly solvable with existing methods. We successfully show SETS automatically discovers a diverse set of optimal behaviors and motion trajectories in real time.
This paper presents a distributed inverse dynamics controller (DIDC) for quadruped robots that addresses the limitations of existing reactive controllers: simplified dynamical models, the inability to handle exact friction cone constraints, and the high computational requirements of whole-body controllers. Current methods either ignore friction constraints entirely or use linear approximations, leading to potential slip and instability, while comprehensive whole-body controllers demand significant computational resources. Our approach uses full rigid-body dynamics and enforces exact friction cone constraints through a novel geometric optimization-based solver. DIDC combines the required generalized forces corresponding to the actuated and unactuated spaces by projecting them onto the actuated space while satisfying the physical constraints and maintaining orthogonality between the base and joint tracking objectives. Experimental validation shows that our approach reduces foot slippage, improves orientation tracking, and converges at least two times faster than existing reactive controllers with generic QP-based implementations. The controller enables stable omnidirectional trotting at various speeds and consumes less power than comparable methods while running efficiently on embedded processors.
In the rapidly advancing realm of visual generation, diffusion models have revolutionized the landscape, marking a significant shift in capabilities with their impressive text-guided generative functions. However, relying solely on text for conditioning these models does not fully cater to the varied and complex requirements of different applications and scenarios. Acknowledging this shortfall, a variety of studies aim to control pre-trained text-to-image (T2I) models to support novel conditions. In this survey, we undertake a thorough review of the literature on controllable generation with T2I diffusion models, covering both the theoretical foundations and practical advancements in this domain. Our review begins with a brief introduction to the basics of denoising diffusion probabilistic models (DDPMs) and widely used T2I diffusion models. We then reveal the controlling mechanisms of diffusion models, theoretically analyzing how novel conditions are introduced into the denoising process for conditional generation. Additionally, we offer a detailed overview of research in this area, organizing it into distinct categories from the condition perspective: generation with specific conditions, generation with multiple conditions, and universal controllable generation. For an exhaustive list of the controllable generation literature surveyed, please refer to our curated repository at \url{//github.com/PRIV-Creation/Awesome-Controllable-T2I-Diffusion-Models}.
注冊地址： 北京市海淀區羊坊店路18號2幢3層301-191