In many practical applications, decision-making processes must balance the costs of acquiring information with the benefits it provides. Traditional control systems often assume full observability, an unrealistic assumption when observations are expensive. We tackle the challenge of simultaneously learning observation and control strategies in such cost-sensitive environments by introducing the Observation-Constrained Markov Decision Process (OCMDP), where the policy influences the observability of the true state. To manage the complexity arising from the combined observation and control actions, we develop an iterative, model-free deep reinforcement learning algorithm that separates the sensing and control components of the policy. This decomposition enables efficient learning in the expanded action space by focusing on when and what to observe, as well as determining optimal control actions, without requiring knowledge of the environment’s dynamics. Experimental results across diverse healthcare tasks and environments demonstrate that our proposed approach substantially reduces observation costs while significantly outperforming baseline methods in efficiency across various real-world scenarios.

This paper formulates the generic Machine Learning (ML) problem into finding the simplest spectral transform form (i.e. one having as many zero coefficients as possible) for an (in)complete binary function. The classical binary logic synthesis problem can be modeled to minimize a single output Boolean function with a two-level structure consisting of an exclusive-OR (EXOR) of ANDs of literals. The innovative approach in this paper is to build and simulate an accelerator that reduces learning to find the exact minimum expression of all 3n Kronecker Reed Muller (KRO) forms of a Boolean function with n input variables. This is in contrast to the previously studied quantum algorithm for the Fixed Polarity Reed-Muller forms (FPRM) which only selects from 2n possible forms. The algorithm, based on repeated application of a ternary Grover's Quantum Search algorithm, was simulated to find the minimum KRO form using a hybrid ternary/binary quantum oracle. This hybrid quantum system was simulated in Matlab and proved to be correct. The method can be also used as a future Quantum EDA Tool for exact minimization of AND/EXOR circuits, including reversible and quantum circuits.

DIGITAL IDEA specializes in creating state-of-the-art visual effects for feature films, both domestic and international. Our slate of films include commercial successes such as Marineboy, Scam, Castaway on the Moon, Take Off, Best Seller, Influence, and Moss. The services we provided for these feature films span over a wide range from digital matte painting, digital environments, to FX simulation, face replacement, and pre-visualization. Digital matte painters created photo-realistic backdrops, virtual environments, and set extensions, our R&D team simulated natural phenomenon such as liquid, fire, smoke, cloth, and fur, and our pre-viz team created animated storyboards to predict and examine characters' movements, angles, camera moves, timing, and focal lengths. The video clip presented at SIGGRAPH ASIA 2010's Computer Animation Festival highlights DIGITAL IDEA's VFX work on feature films released since 2009. Here we show before and after shots that illustrate how a director's imagination becomes a reality with the right tools and right talent: how a previsualization shapes the final shot, and how impossible stunts and camera moves are made possible with the aid of computer graphics.

Due to the lack of semantic descriptions of the Web services, the search results returned by the service registries are effectively inadequate. This paper presents the Semantic Web services Clustering (SWSC) method that extends the semantic representation of services and groups the similar Web services in order to improve the service discovery. The empirical analysis shows the improvement in service discovery with the use of SWSC.