Altruistic learning.

The origin of altruism remains one of the most enduring puzzles of human behaviour. Indeed, true altruism is often thought either not to exist, or to arise merely as a miscalculation of otherwise selfish behaviour. In this paper, we argue that altruism emerges directly from the way in which distinct human decision-making systems learn about rewards. Using insights provided by neurobiological accounts of human decision-making, we suggest that reinforcement learning in game-theoretic social interactions (habitisation over either individuals or games) and observational learning (either imitative of inference based) lead to altruistic behaviour. This arises not only as a result of computational efficiency in the face of processing complexity, but as a direct consequence of optimal inference in the face of uncertainty. Critically, we argue that the fact that evolutionary pressure acts not over the object of learning ('what' is learned), but over the learning systems themselves ('how' things are learned), enables the evolution of altruism despite the direct threat posed by free-riders.

On-line policy optimisation of spoken dialogue systems via live interaction with human subjects

Statistical dialogue models have required a large number of dialogues to optimise the dialogue policy, relying on the use of a simulated user. This results in a mismatch between training and live conditions, and significant development costs for the simulator thereby mitigating many of the claimed benefits of such models. Recent work on Gaussian process reinforcement learning, has shown that learning can be substantially accelerated. This paper reports on an experiment to learn a policy for a real-world task directly from human interaction using rewards provided by users. It shows that a usable policy can be learnt in just a few hundred dialogues without needing a user simulator and, using a learning strategy that reduces the risk of taking bad actions. The paper also investigates adaptation behaviour when the system continues learning for several thousand dialogues and highlights the need for robustness to noisy rewards. © 2011 IEEE.

Interfacial recognition of human prostate-specific antigen by immobilized monoclonal antibody: effects of solution conditions and surface chemistry.

The specific recognition between monoclonal antibody (anti-human prostate-specific antigen, anti-hPSA) and its antigen (human prostate-specific antigen, hPSA) has promising applications in prostate cancer diagnostics and other biosensor applications. However, because of steric constraints associated with interfacial packing and molecular orientations, the binding efficiency is often very low. In this study, spectroscopic ellipsometry and neutron reflection have been used to investigate how solution pH, salt concentration and surface chemistry affect antibody adsorption and subsequent antigen binding. The adsorbed amount of antibody was found to vary with pH and the maximum adsorption occurred between pH 5 and 6, close to the isoelectric point of the antibody. By contrast, the highest antigen binding efficiency occurred close to the neutral pH. Increasing the ionic strength reduced antibody adsorbed amount at the silica-water interface but had little effect on antigen binding. Further studies of antibody adsorption on hydrophobic C8 (octyltrimethoxysilane) surface and chemical attachment of antibody on (3-mercaptopropyl)trimethoxysilane/4-maleimidobutyric acid N-hydroxysuccinimide ester-modified surface have also been undertaken. It was found that on all surfaces studied, the antibody predominantly adopted the 'flat on' orientation, and antigen-binding capabilities were comparable. The results indicate that antibody immobilization via appropriate physical adsorption can replace elaborate interfacial molecular engineering involving complex covalent attachments.