<p dir="ltr">In the context of reinforcement learning, extensive research has shown how reinforcement learning was facilitated by the estimation of uncertainty to improve the ability to make decisions. However, the constraints imposed by the limited observation of the process of forming environment representation have seldom been a subject of discussion. Thus, the study intended to demonstrate that when incorporating a limited memory into uncertainty estimation, individuals potentially misestimate outcomes and environmental statistics. The study included a computational model that included the process of active working memory and lateral inhibition in working memory (WM) to describe how the relevant information was chosen and stored to form estimations of uncertainty in forming outcome expectations. The active working memory maintained relevant information not just by the recent memory, but also with utility. With the relevant information stored in WM, the model was able to estimate expected uncertainty and perceived volatility and detect contextual changes or dynamics in the outcome structure. Two experiments to investigate limitations in information availability and uncertainty estimation were carried out. The first experiment investigated the impact of cognitive loading on the reliance on memories to form outcome estimation. The findings revealed that introducing cognitive loading diminished the reliance on memory for uncertainty estimations and lowered the expected uncertainty, leading to an increased perception of environmental volatility. The second experiment investigated the ability to detect changes in outcome noise under different conditions of outcome exposure. The study found differences in the mechanisms used for detecting environmental changes in various conditions. Through the experiments and model fitting, the study showed that the misestimation of uncertainties was reliant on individual experiences and relevant information stored in WM under a limited capacity.</p>

No description available.

No description available.

No description available.

No description available.

No description available.

No description available.

Multipath on far-field antenna ranges causes distortion of antenna pattern measurements: The multi path components have differing path lengths and hence can be separated by illuminating the antenna under test with short-duration pulses. Alternatively, antenna measurements can be made in the frequency domain, and the Fourier transform can be used to relate the frequency-domain measurements to the antenna's time-domain response. The interference then can be removed with a time-domain gate, and transformed back into the frequency domain to yield improved CW antenna patterns. Virginia Tech has recently completed a major upgrade of their far-field antenna range and implemented a system to perform this data collection and data processing. This thesis describes the principles and implementation of the time-domain processing part of the system. Further, it demonstrates the validity of the method by showing the improvements in pattern measurement that have been achieved with the new system. / Master of Science

As transistor sizes scale down and levels of integration increase, leakage power has become a critical problem in modern low-power microprocessors. This is especially true for ultra-low-voltage (ULV) circuits, where high levels of leakage force designers to chose relatively high threshold voltages, which limits performance. In this thesis, an industry-standard technique known as power-gating is explored, whereby transistors are used to disconnect the power from idle portions of a chip. Present power-gating implementations suffer from limitations including non-zero off-state leakage, which can aggregate to a large amount of wasted energy during long idle periods, and high energy overhead, which limits its use to long-term system-wide sleep modes. As this thesis will show however, by vastly increasing the effectiveness of power-gating through the use of emerging technologies, and by implementing aggressive hardware-oriented power-gating policies, leakage in microprocessors can be eliminated to a large extent. This allows the threshold voltage to be lowered, leading to ULV microprocessors with both low switching energy and high performance. The first emerging technology investigated is the Nanoelectromechnical-Systems (NEMS) switch, which is a CMOS-compatible mechanical relay with near-infinite off-resistance and low on-resistance. When used for power-gating, this switch completely eliminates off-state leakage, yet is compact enough to be contained on die. This has tremendous benefits for applications with long sleep times. For example, a NEMS-power-gated architecture performing an FFT per hour consumes 30 times less power than a transistor-power-gated architecture. Additionally, the low on-resistance can lower power-gating area overhead by 36-83\%. The second technology targets the high energy overhead associated with powering a circuit on and off. This thesis demonstrates that a new logic style specifically designed for ULV operation, Sense Amplifier Pass Transistor Logic (SAPTL), requires power-gates that are 8-10 times smaller, and consumes up to 15 times less boot-up energy, compared to static-CMOS. These abilities enable effective power-gating of an SAPTL circuit, even for very short idle periods. Microprocessor simulations demonstrate that a fine-grained power-gating policy, along with this drastically lower overhead, can result in up to a 44\% drop in energy. Encompassing these investigations is an energy estimation framework built around a cycle-accurate microprocessor simulator, which allows a wide range of circuit and power-gating parameters to be optimized. This framework implements two hardware-based power-gating schedulers that are completely invisible to the OS, and have extremely low hardware overhead, allowing for a large number of power-gated regions. All together, this thesis represents the most complete and forward-looking study on power-gating in the ULV region. The results demonstrate that aggressive power-gating allows designers to leverage the very low switching energy of ULV operation, while achieving performance levels that can greatly expand the capabilities of energy-constrained systems. / Ph. D.

Yes / In a previous study we identified an extensive gating network within the inwardly rectifying Kir1.1 (ROMK) channel by combining systematic scanning mutagenesis and functional analysis with structural models of the channel in the closed, pre-open and open states. This extensive network appeared to stabilize the open and pre-open states, but the network fragmented upon channel closure. In this study we have analyzed the gating kinetics of different mutations within key parts of this gating network. These results suggest that the structure of the transition state (TS), which connects the pre-open and closed states of the channel, more closely resembles the structure of the pre-open state. Furthermore, the G-loop, which occurs at the center of this extensive gating network, appears to become unstructured in the TS because mutations within this region have a 'catalytic' effect upon the channel gating kinetics. / Deutsche Forschungsgemeinschaft, the Wellcome Trust (083547/ Z/07/Z and 092970/Z/10/Z) and the British Heart Foundation (PG/09/016/ 26992).