Search

Phenology, the timing of cyclical plant life events such as leaf emergence and coloration, is crucial in the bio-climatic system. Climate change drives shifts in these phenological events, impacting ecosystems and the climate itself. Accurate phenology models are essential to predict the occurrence of these phases under changing climatic conditions. Existing methods include hypothesis-driven process models and data-driven statistical approaches. Process models account for dormancy stages and various phenology drivers, while statistical models typically rely on linear or traditional machine learning techniques. Research shows that process models often outperform statistical methods when predicting under climate conditions outside historical ranges, especially with climate change scenarios. However, deep learning approaches remain underexplored in climate phenology modeling. We introduce PhenoFormer, a neural architecture better suited than traditional statistical methods at predicting phenology under shift in climate data distribution, while also bringing significant improvements or performing on par to the best performing process-based models. Our numerical experiments on a 70-year dataset of 70,000 phenological observations from 9 woody species in Switzerland show that PhenoFormer outperforms traditional machine learning methods by an average of 13% R2 and 1.1 days RMSE for spring phenology, and 11% R2 and 0.7 days RMSE for autumn phenology, while matching or exceeding the best process-based models. Our results demonstrate that deep learning has the potential to be a valuable methodological tool for accurate climate-phenology prediction, and our PhenoFormer is a first promising step in improving phenological predictions before a complete understanding of the underlying physiological mechanisms is available., Comment: journal-preprint

Molecular photoabsorption results in an electronic excitation/ionization which couples to the rearrangement of the nuclei. The resulting intertwined change of nuclear and electronic degrees of freedom determines the conversion of photoenergy into other molecular energy forms. Nucleobases are excellent candidates for studying such dynamics, and great effort has been taken in the past to observe the electronic changes induced by the initial excitation in a time-resolved manner using ultrafast electron spectroscopy. The linked geometrical changes during nucleobase photorelaxation have so far not been observed directly in time-resolved experiments. Here, we present a study on a thionucleobase, where we extract comprehensive information on the molecular rearrangement using Coulomb explosion imaging. Our measurement links the extracted deplanarization of the molecular geometry to the previously studied temporal evolution of the electronic properties of the system. In particular, the protons of the exploded molecule are well-suited messengers carrying rich information on the molecule's geometry at distinct times after the initial electronic excitation. The combination of ultrashort laser pulses to trigger molecular dynamics, intense X-ray free-electron laser pulses for the explosion of the molecule, and multi-particle coincidence detection opens new avenues for time-resolved studies of complex molecules in the gas phase., Comment: 19 pages, 8 figures

We apply multi-agent deep reinforcement learning (RL) to train end-to-end robot soccer policies with fully onboard computation and sensing via egocentric RGB vision. This setting reflects many challenges of real-world robotics, including active perception, agile full-body control, and long-horizon planning in a dynamic, partially-observable, multi-agent domain. We rely on large-scale, simulation-based data generation to obtain complex behaviors from egocentric vision which can be successfully transferred to physical robots using low-cost sensors. To achieve adequate visual realism, our simulation combines rigid-body physics with learned, realistic rendering via multiple Neural Radiance Fields (NeRFs). We combine teacher-based multi-agent RL and cross-experiment data reuse to enable the discovery of sophisticated soccer strategies. We analyze active-perception behaviors including object tracking and ball seeking that emerge when simply optimizing perception-agnostic soccer play. The agents display equivalent levels of performance and agility as policies with access to privileged, ground-truth state. To our knowledge, this paper constitutes a first demonstration of end-to-end training for multi-agent robot soccer, mapping raw pixel observations to joint-level actions, that can be deployed in the real world. Videos of the game-play and analyses can be seen on our website https://sites.google.com/view/vision-soccer .

The challenge of accessing historical patient data for clinical research, while adhering to privacy regulations, is a significant obstacle in medical science. An innovative approach to circumvent this issue involves utilising synthetic medical records that mirror real patient data without compromising individual privacy. The creation of these synthetic datasets, particularly without using actual patient data to train Large Language Models (LLMs), presents a novel solution as gaining access to sensitive patient information to train models is also a challenge. This study assesses the capability of the Llama 2 LLM to create synthetic medical records that accurately reflect real patient information, employing zero-shot and few-shot prompting strategies for comparison against fine-tuned methodologies that do require sensitive patient data during training. We focus on generating synthetic narratives for the History of Present Illness section, utilising data from the MIMIC-IV dataset for comparison. In this work introduce a novel prompting technique that leverages a chain-of-thought approach, enhancing the model's ability to generate more accurate and contextually relevant medical narratives without prior fine-tuning. Our findings suggest that this chain-of-thought prompted approach allows the zero-shot model to achieve results on par with those of fine-tuned models, based on Rouge metrics evaluation., Comment: 4 pages

We investigate whether Deep Reinforcement Learning (Deep RL) is able to synthesize sophisticated and safe movement skills for a low-cost, miniature humanoid robot that can be composed into complex behavioral strategies in dynamic environments. We used Deep RL to train a humanoid robot with 20 actuated joints to play a simplified one-versus-one (1v1) soccer game. The resulting agent exhibits robust and dynamic movement skills such as rapid fall recovery, walking, turning, kicking and more; and it transitions between them in a smooth, stable, and efficient manner. The agent's locomotion and tactical behavior adapts to specific game contexts in a way that would be impractical to manually design. The agent also developed a basic strategic understanding of the game, and learned, for instance, to anticipate ball movements and to block opponent shots. Our agent was trained in simulation and transferred to real robots zero-shot. We found that a combination of sufficiently high-frequency control, targeted dynamics randomization, and perturbations during training in simulation enabled good-quality transfer. Although the robots are inherently fragile, basic regularization of the behavior during training led the robots to learn safe and effective movements while still performing in a dynamic and agile way -- well beyond what is intuitively expected from the robot. Indeed, in experiments, they walked 181% faster, turned 302% faster, took 63% less time to get up, and kicked a ball 34% faster than a scripted baseline, while efficiently combining the skills to achieve the longer term objectives., Comment: Project website: https://sites.google.com/view/op3-soccer

We introduce the use of generative adversarial learning to compute equilibria in general game-theoretic settings, specifically the generalized Nash equilibrium (GNE) in pseudo-games, and its specific instantiation as the competitive equilibrium (CE) in Arrow-Debreu competitive economies. Pseudo-games are a generalization of games in which players' actions affect not only the payoffs of other players but also their feasible action spaces. Although the computation of GNE and CE is intractable in the worst-case, i.e., PPAD-hard, in practice, many applications only require solutions with high accuracy in expectation over a distribution of problem instances. We introduce Generative Adversarial Equilibrium Solvers (GAES): a family of generative adversarial neural networks that can learn GNE and CE from only a sample of problem instances. We provide computational and sample complexity bounds, and apply the framework to finding Nash equilibria in normal-form games, CE in Arrow-Debreu competitive economies, and GNE in an environmental economic model of the Kyoto mechanism., Comment: 41 pages, 13 figures

The Game Theory & Multi-Agent team at DeepMind studies several aspects of multi-agent learning ranging from computing approximations to fundamental concepts in game theory to simulating social dilemmas in rich spatial environments and training 3-d humanoids in difficult team coordination tasks. A signature aim of our group is to use the resources and expertise made available to us at DeepMind in deep reinforcement learning to explore multi-agent systems in complex environments and use these benchmarks to advance our understanding. Here, we summarise the recent work of our team and present a taxonomy that we feel highlights many important open challenges in multi-agent research., Comment: Published in AI Communications 2022

We present real-world data processing on measured electron time-of-flight data via neural networks. Specifically, the use of disentangled variational autoencoders on data from a diagnostic instrument for online wavelength monitoring at the free electron laser FLASH in Hamburg. Without a-priori knowledge the network is able to find representations of single-shot FEL spectra, which have a low signal-to-noise ratio. This reveals, in a directly human-interpretable way, crucial information about the photon properties. The central photon energy and the intensity as well as very detector-specific features are identified. The network is also capable of data cleaning, i.e. denoising, as well as the removal of artefacts. In the reconstruction, this allows for identification of signatures with very low intensity which are hardly recognisable in the raw data. In this particular case, the network enhances the quality of the diagnostic analysis at FLASH. However, this unsupervised method also has the potential to improve the analysis of other similar types of spectroscopy data.

Intelligent behaviour in the physical world exhibits structure at multiple spatial and temporal scales. Although movements are ultimately executed at the level of instantaneous muscle tensions or joint torques, they must be selected to serve goals defined on much longer timescales, and in terms of relations that extend far beyond the body itself, ultimately involving coordination with other agents. Recent research in artificial intelligence has shown the promise of learning-based approaches to the respective problems of complex movement, longer-term planning and multi-agent coordination. However, there is limited research aimed at their integration. We study this problem by training teams of physically simulated humanoid avatars to play football in a realistic virtual environment. We develop a method that combines imitation learning, single- and multi-agent reinforcement learning and population-based training, and makes use of transferable representations of behaviour for decision making at different levels of abstraction. In a sequence of stages, players first learn to control a fully articulated body to perform realistic, human-like movements such as running and turning; they then acquire mid-level football skills such as dribbling and shooting; finally, they develop awareness of others and play as a team, bridging the gap between low-level motor control at a timescale of milliseconds, and coordinated goal-directed behaviour as a team at the timescale of tens of seconds. We investigate the emergence of behaviours at different levels of abstraction, as well as the representations that underlie these behaviours using several analysis techniques, including statistics from real-world sports analytics. Our work constitutes a complete demonstration of integrated decision-making at multiple scales in a physically embodied multi-agent setting. See project video at https://youtu.be/KHMwq9pv7mg.

The conversion of photon energy into other energetic forms in molecules is accompanied by charge moving on ultrafast timescales. We directly observe the charge motion at a specific site in an electronically excited molecule using time-resolved x-ray photoelectron spectroscopy (TR-XPS). We extend the concept of static chemical shift from conventional XPS by the excited-state chemical shift (ESCS), which is connected to the charge in the framework of a potential model. This allows us to invert TR-XPS spectra to the dynamic charge at a specific atom. We demonstrate the power of TR-XPS by using sulphur 2p-core-electron-emission probing to study the UV-excited dynamics of 2-thiouracil. The new method allows us to discover that a major part of the population relaxes to the molecular ground state within 220-250 fs. In addition, a 250-fs oscillation, visible in the kinetic energy of the TR-XPS, reveals a coherent exchange of population among electronic states.

We study the problem of emergent communication, in which language arises because speakers and listeners must communicate information in order to solve tasks. In temporally extended reinforcement learning domains, it has proved hard to learn such communication without centralized training of agents, due in part to a difficult joint exploration problem. We introduce inductive biases for positive signalling and positive listening, which ease this problem. In a simple one-step environment, we demonstrate how these biases ease the learning problem. We also apply our methods to a more extended environment, showing that agents with these inductive biases achieve better performance, and analyse the resulting communication protocols., Comment: Accepted at NeurIPS 2019

Rare diseases affect hundreds of millions of people worldwide but are hard to detect since they have extremely low prevalence rates (varying from 1/1,000 to 1/200,000 patients) and are massively underdiagnosed. How do we reliably detect rare diseases with such low prevalence rates? How to further leverage patients with possibly uncertain diagnosis to improve detection? In this paper, we propose a Complementary pattern Augmentation (CONAN) framework for rare disease detection. CONAN combines ideas from both adversarial training and max-margin classification. It first learns self-attentive and hierarchical embedding for patient pattern characterization. Then, we develop a complementary generative adversarial networks (GAN) model to generate candidate positive and negative samples from the uncertain patients by encouraging a max-margin between classes. In addition, CONAN has a disease detector that serves as the discriminator during the adversarial training for identifying rare diseases. We evaluated CONAN on two disease detection tasks. For low prevalence inflammatory bowel disease (IBD) detection, CONAN achieved .96 precision recall area under the curve (PR-AUC) and 50.1% relative improvement over best baseline. For rare disease idiopathic pulmonary fibrosis (IPF) detection, CONAN achieves .22 PR-AUC with 41.3% relative improvement over the best baseline.

This paper investigates a population-based training regime based on game-theoretic principles called Policy-Spaced Response Oracles (PSRO). PSRO is general in the sense that it (1) encompasses well-known algorithms such as fictitious play and double oracle as special cases, and (2) in principle applies to general-sum, many-player games. Despite this, prior studies of PSRO have been focused on two-player zero-sum games, a regime wherein Nash equilibria are tractably computable. In moving from two-player zero-sum games to more general settings, computation of Nash equilibria quickly becomes infeasible. Here, we extend the theoretical underpinnings of PSRO by considering an alternative solution concept, $\alpha$-Rank, which is unique (thus faces no equilibrium selection issues, unlike Nash) and applies readily to general-sum, many-player settings. We establish convergence guarantees in several games classes, and identify links between Nash equilibria and $\alpha$-Rank. We demonstrate the competitive performance of $\alpha$-Rank-based PSRO against an exact Nash solver-based PSRO in 2-player Kuhn and Leduc Poker. We then go beyond the reach of prior PSRO applications by considering 3- to 5-player poker games, yielding instances where $\alpha$-Rank achieves faster convergence than approximate Nash solvers, thus establishing it as a favorable general games solver. We also carry out an initial empirical validation in MuJoCo soccer, illustrating the feasibility of the proposed approach in another complex domain.

We study the emergence of cooperative behaviors in reinforcement learning agents by introducing a challenging competitive multi-agent soccer environment with continuous simulated physics. We demonstrate that decentralized, population-based training with co-play can lead to a progression in agents' behaviors: from random, to simple ball chasing, and finally showing evidence of cooperation. Our study highlights several of the challenges encountered in large scale multi-agent training in continuous control. In particular, we demonstrate that the automatic optimization of simple shaping rewards, not themselves conducive to co-operative behavior, can lead to long-horizon team behavior. We further apply an evaluation scheme, grounded by game theoretic principals, that can assess agent performance in the absence of pre-defined evaluation tasks or human baselines.

Recent progress in artificial intelligence through reinforcement learning (RL) has shown great success on increasingly complex single-agent environments and two-player turn-based games. However, the real-world contains multiple agents, each learning and acting independently to cooperate and compete with other agents, and environments reflecting this degree of complexity remain an open challenge. In this work, we demonstrate for the first time that an agent can achieve human-level in a popular 3D multiplayer first-person video game, Quake III Arena Capture the Flag, using only pixels and game points as input. These results were achieved by a novel two-tier optimisation process in which a population of independent RL agents are trained concurrently from thousands of parallel matches with agents playing in teams together and against each other on randomly generated environments. Each agent in the population learns its own internal reward signal to complement the sparse delayed reward from winning, and selects actions using a novel temporally hierarchical representation that enables the agent to reason at multiple timescales. During game-play, these agents display human-like behaviours such as navigating, following, and defending based on a rich learned representation that is shown to encode high-level game knowledge. In an extensive tournament-style evaluation the trained agents exceeded the win-rate of strong human players both as teammates and opponents, and proved far stronger than existing state-of-the-art agents. These results demonstrate a significant jump in the capabilities of artificial agents, bringing us closer to the goal of human-level intelligence.

Polarized light microscopy, as a contrast-enhancing technique for optically anisotropic materials, is a method well suited for the investigation of a wide variety of effects in solid-state physics, as for example birefringence in crystals or the magneto-optical Kerr effect (MOKE). We present a microscopy setup that combines a widefield microscope and a confocal scanning laser microscope with polarization-sensitive detectors. By using a high numerical aperture objective, a spatial resolution of about 240 nm at a wavelength of 405 nm is achieved. The sample is mounted on a $^4$He continuous flow cryostat providing a temperature range between 4 K and 300 K, and electromagnets are used to apply magnetic fields of up to 800 mT with variable in-plane orientation and 20 mT with out-of-plane orientation. Typical applications of the polarizing microscope are the imaging of the in-plane and out-of-plane magnetization via the longitudinal and polar MOKE, imaging of magnetic flux structures in superconductors covered with a magneto-optical indicator film via Faraday effect or imaging of structural features, such as twin-walls in tetragonal SrTiO$_3$. The scanning laser microscope furthermore offers the possibility to gain local information on electric transport properties of a sample by detecting the beam-induced voltage change across a current-biased sample. This combination of magnetic, structural and electric imaging capabilities makes the microscope a viable tool for research in the fields of oxide electronics, spintronics, magnetism and superconductivity., Comment: 14 pages, 11 figures. The following article has been accepted by Review of Scientific Instruments. After it is published, it will be found at http://aip.scitation.org/journal/rsi

Domain squatting is a common adversarial practice where attackers register domain names that are purposefully similar to popular domains. In this work, we study a specific type of domain squatting called "combosquatting," in which attackers register domains that combine a popular trademark with one or more phrases (e.g., betterfacebook[.]com, youtube-live[.]com). We perform the first large-scale, empirical study of combosquatting by analyzing more than 468 billion DNS records---collected from passive and active DNS data sources over almost six years. We find that almost 60% of abusive combosquatting domains live for more than 1,000 days, and even worse, we observe increased activity associated with combosquatting year over year. Moreover, we show that combosquatting is used to perform a spectrum of different types of abuse including phishing, social engineering, affiliate abuse, trademark abuse, and even advanced persistent threats. Our results suggest that combosquatting is a real problem that requires increased scrutiny by the security community., Comment: ACM CCS 17

Quantum steering describes how local actions on a quantum system can affect another, space-like separated, quantum state. Lately, quantum steering has been formulated also for time-like scenarios and for quantum channels. We approach all the three scenarios as one using tools from Stinespring dilations of quantum channels. By applying our technique we link all three steering problems one-to-one with the incompatibility of quantum measurements, a result formerly known only for spatial steering. We exploit this connection by showing how measurement uncertainty relations can be used as tight steering inequalities for all three scenarios. Moreover, we show that certain notions of temporal and spatial steering are fully equivalent and prove a hierarchy between temporal steering and macrorealistic hidden variable models., Comment: 7 pages, 2 figures

We study the problem of cooperative multi-agent reinforcement learning with a single joint reward signal. This class of learning problems is difficult because of the often large combined action and observation spaces. In the fully centralized and decentralized approaches, we find the problem of spurious rewards and a phenomenon we call the "lazy agent" problem, which arises due to partial observability. We address these problems by training individual agents with a novel value decomposition network architecture, which learns to decompose the team value function into agent-wise value functions. We perform an experimental evaluation across a range of partially-observable multi-agent domains and show that learning such value-decompositions leads to superior results, in particular when combined with weight sharing, role information and information channels.

We present a unifying framework for adapting the update direction in gradient-based iterative optimization methods. As natural special cases we re-derive classical momentum and Nesterov's accelerated gradient method, lending a new intuitive interpretation to the latter algorithm. We show that a new algorithm, which we term Regularised Gradient Descent, can converge more quickly than either Nesterov's algorithm or the classical momentum algorithm.

Approximate Newton methods are a standard optimization tool which aim to maintain the benefits of Newton's method, such as a fast rate of convergence, whilst alleviating its drawbacks, such as computationally expensive calculation or estimation of the inverse Hessian. In this work we investigate approximate Newton methods for policy optimization in Markov Decision Processes (MDPs). We first analyse the structure of the Hessian of the objective function for MDPs. We show that, like the gradient, the Hessian exhibits useful structure in the context of MDPs and we use this analysis to motivate two Gauss-Newton Methods for MDPs. Like the Gauss-Newton method for non-linear least squares, these methods involve approximating the Hessian by ignoring certain terms in the Hessian which are difficult to estimate. The approximate Hessians possess desirable properties, such as negative definiteness, and we demonstrate several important performance guarantees including guaranteed ascent directions, invariance to affine transformation of the parameter space, and convergence guarantees. We finally provide a unifying perspective of key policy search algorithms, demonstrating that our second Gauss-Newton algorithm is closely related to both the EM-algorithm and natural gradient ascent applied to MDPs, but performs significantly better in practice on a range of challenging domains.

Recently two approximate Newton methods were proposed for the optimisation of Markov Decision Processes. While these methods were shown to have desirable properties, such as a guarantee that the preconditioner is negative-semidefinite when the policy is $\log$-concave with respect to the policy parameters, and were demonstrated to have strong empirical performance in challenging domains, such as the game of Tetris, no convergence analysis was provided. The purpose of this paper is to provide such an analysis. We start by providing a detailed analysis of the Hessian of a Markov Decision Process, which is formed of a negative-semidefinite component, a positive-semidefinite component and a remainder term. The first part of our analysis details how the negative-semidefinite and positive-semidefinite components relate to each other, and how these two terms contribute to the Hessian. The next part of our analysis shows that under certain conditions, relating to the richness of the policy class, the remainder term in the Hessian vanishes in the vicinity of a local optimum. Finally, we bound the behaviour of this remainder term in terms of the mixing time of the Markov chain induced by the policy parameters, where this part of the analysis is applicable over the entire parameter space. Given this analysis of the Hessian we then provide our local convergence analysis of the approximate Newton framework., Comment: This work has been removed because a more recent piece (A Gauss-Newton method for Markov Decision Processes, T. Furmston & G. Lever) of work has subsumed it

We report the observation of the quantum-confined Stark effect in Ge/SiGe multiple quantum well heterostructures grown on Si(0.22)Ge(0.78) virtual substrates. The large compressive strain in the Ge quantum well layers caused by the lattice mismatch with the virtual substrate results in a blue shift of the direct absorption edge, as well as a reduction in the \Gamma-valley scattering lifetime because of strain-induced splittings of the conduction band valleys. We investigate theoretically the \Gamma-valley carrier lifetimes by evaluating the \Gamma-to-L and \Gamma-to-\Delta{} scattering rates in strained Ge/SiGe semiconductor heterostructures. These scattering rates are used to determine the lifetime broadening of excitonic peaks and the indirect absorption in simulated absorption spectra, which are compared with measured absorption spectra for quantum well structures with systematically-varied dimensions. We find that \Gamma-to-\Delta{} scattering is significant in compressively strained Ge quantum wells and that the \Gamma-valley electron lifetime is less than 50 fs in the highly-strained structures reported here, where \Gamma-to-\Delta{} scattering accounted for approximately half of the total scattering rate., Comment: 8 pages, 8 figures

A detailed study of energy differences between the highest occupied and lowest unoccupied molecular orbitals (HOMO-LUMO gaps) in protein systems and water clusters is presented. Recent work questioning the applicability of Kohn-Sham density-functional theory to proteins and large water clusters (E. Rudberg, J. Phys.: Condens. Mat. 2012, 24, 072202) has demonstrated vanishing HOMO-LUMO gaps for these systems, which is generally attributed to the treatment of exchange in the functional used. The present work shows that the vanishing gap is, in fact, an electrostatic artefact of the method used to prepare the system. Practical solutions for ensuring the gap is maintained when the system size is increased are demonstrated. This work has important implications for the use of large-scale density-functional theory in biomolecular systems, particularly in the simulation of photoemission, optical absorption and electronic transport, all of which depend critically on differences between energies of molecular orbitals., Comment: 13 pages, 4 figures

We propose a new, nonparametric approach to learning and representing transition dynamics in Markov decision processes (MDPs), which can be combined easily with dynamic programming methods for policy optimisation and value estimation. This approach makes use of a recently developed representation of conditional distributions as \emph{embeddings} in a reproducing kernel Hilbert space (RKHS). Such representations bypass the need for estimating transition probabilities or densities, and apply to any domain on which kernels can be defined. This avoids the need to calculate intractable integrals, since expectations are represented as RKHS inner products whose computation has linear complexity in the number of points used to represent the embedding. We provide guarantees for the proposed applications in MDPs: in the context of a value iteration algorithm, we prove convergence to either the optimal policy, or to the closest projection of the optimal policy in our model class (an RKHS), under reasonable assumptions. In experiments, we investigate a learning task in a typical classical control setting (the under-actuated pendulum), and on a navigation problem where only images from a sensor are observed. For policy optimisation we compare with least-squares policy iteration where a Gaussian process is used for value function estimation. For value estimation we also compare to the NPDP method. Our approach achieves better performance in all experiments., Comment: ICML2012

Silicon-based terahertz quantum cascade lasers (QCLs) offer potential advantages over existing III-V devices. Although coherent electron transport effects are known to be important in QCLs, they have never been considered in Si-based device designs. We describe a density matrix transport model that is designed to be more general than those in previous studies and to require less a priori knowlege of electronic bandstructure, allowing its use in semi-automated design procedures. The basis of the model includes all states involved in interperiod transport, and our steady-state solution extends beyond the rotating-wave approximation by including DC and counter-propagating terms. We simulate the potential performance of bound-to-continuum Ge/SiGe QCLs and find that devices with 4-5-nm-thick barriers give the highest simulated optical gain. We also examine the effects of interdiffusion between Ge and SiGe layers; we show that if it is taken into account in the design, interdiffusion lengths of up to 1.5 nm do not significantly affect the simulated device performance., Comment: 8 pages, 4 figures

We demonstrate an equivalence between reproducing kernel Hilbert space (RKHS) embeddings of conditional distributions and vector-valued regressors. This connection introduces a natural regularized loss function which the RKHS embeddings minimise, providing an intuitive understanding of the embeddings and a justification for their use. Furthermore, the equivalence allows the application of vector-valued regression methods and results to the problem of learning conditional distributions. Using this link we derive a sparse version of the embedding by considering alternative formulations. Further, by applying convergence results for vector-valued regression to the embedding problem we derive minimax convergence rates which are O(\log(n)/n) -- compared to current state of the art rates of O(n^{-1/4}) -- and are valid under milder and more intuitive assumptions. These minimax upper rates coincide with lower rates up to a logarithmic factor, showing that the embedding method achieves nearly optimal rates. We study our sparse embedding algorithm in a reinforcement learning task where the algorithm shows significant improvement in sparsity over an incomplete Cholesky decomposition.

A new catalogue of extrasolar planets is presented by re-analysing a selection of published radial velocity data sets using EXOFIT (Balan & Lahav 2009). All objects are treated on an equal footing within a Bayesian framework, to give orbital parameters for 94 exoplanetary systems. Model selection (between one- and two-planet solutions) is then performed, using both a visual flagging method and a standard chi-square analysis, with agreement between the two methods for 99% of the systems considered. The catalogue is to be made available online, and this 'proof of concept' study may be maintained and extended in the future to incorporate all systems with publicly available radial velocity data, as well as transit and microlensing data., Comment: 16 pages, 6 figures, 6 tables

We propose a new, nonparametric approach to estimating the value function in reinforcement learning. This approach makes use of a recently developed representation of conditional distributions as functions in a reproducing kernel Hilbert space. Such representations bypass the need for estimating transition probabilities, and apply to any domain on which kernels can be defined. Our approach avoids the need to approximate intractable integrals since expectations are represented as RKHS inner products whose computation has linear complexity in the sample size. Thus, we can efficiently perform value function estimation in a wide variety of settings, including finite state spaces, continuous states spaces, and partially observable tasks where only sensor measurements are available. A second advantage of the approach is that we learn the conditional distribution representation from a training sample, and do not require an exhaustive exploration of the state space. We prove convergence of our approach either to the optimal policy, or to the closest projection of the optimal policy in our model class, under reasonable assumptions. In experiments, we demonstrate the performance of our algorithm on a learning task in a continuous state space (the under-actuated pendulum), and on a navigation problem where only images from a sensor are observed. We compare with least-squares policy iteration where a Gaussian process is used for value function estimation. Our algorithm achieves better performance in both tasks., Comment: The article can now be found under arXiv:1206.4655. We combined both versions and are withdrawing this version because of the resulting redundancy

We propose a method to efficiently construct data-dependent kernels which can make use of large quantities of (unlabeled) data. Our construction makes an approximation in the standard construction of semi-supervised kernels in Sindhwani et al. 2005. In typical cases these kernels can be computed in nearly-linear time (in the amount of data), improving on the cubic time of the standard construction, enabling large scale semi-supervised learning in a variety of contexts. The methods are validated on semi-supervised and unsupervised problems on data sets containing upto 64,000 sample points.

Silicon-based quantum cascade lasers (QCLs) offer the prospect of integrating coherent THz radiation sources with silicon microelectronics. Theoretical studies have proposed a variety of n-type SiGe-based heterostructures as design candidates, however the optimal material configuration remains unclear. In this work, an optimization algorithm is used to design equivalent THz QCLs in three recently-proposed configurations [(001) Ge/GeSi, (001) Si/SiGe and (111) Si/SiGe], with emission frequencies of 3 and 4 THz. A systematic comparison of the electronic and optical properties is presented. A semi-classical electron transport simulation is used to model the charge carrier dynamics and calculate the peak gain, the corresponding current density and the maximum operating temperature. It is shown that (001) Ge/GeSi structures yield the best simulated performance at both emission frequencies., Comment: 9 pages, 7 figures

Although most work toward the realization of group IV quantum cascade lasers (QCLs) has focused on valence-band transitions, there are many desirable properties associated with the conduction band. We show that the commonly cited shortcomings of n-type Si/SiGe heterostructures can be overcome by moving to the (111) growth direction. Specifically, a large band offset and low effective mass are achievable and subband degeneracy is preserved. We predict net gain up to lattice temperatures of 90 K in a bound-to-continuum QCL with a double-metal waveguide, and show that a Ge interdiffusion length of at least 8 angstroms across interfaces is tolerable., Comment: 7 pages, 6 figures

Based on descriptions submitted to a national panel on exemplary Distance Education (DE) programs, this monograph describes state-of-the-art DE programs at 16 community colleges. First, an introduction reviews the history of DE systems and DE in community colleges, describes the importance of revising curricula for distance delivery, discusses materials and technologies associated with DE, and reviews characteristics of synchronous and asynchronous delivery systems. The subsequent chapters then describe programs at the following community colleges: Austin Community College (Texas), Chattanooga State Technical Community College (Tennessee), Chemeketa Community College (Oregon), Dallas County Community College District (Texas), Florida Community College at Jacksonville, Genesee Community College (New York), Kirkwood Community College (Iowa), Miami-Dade Community College District (Florida), Northern Virginia Community College, Northwestern Michigan College, Rio Salado College (Arizona), San Diego City College (California), Sinclair Community College (Ohio), Tarrant County Junior College District (Texas), University College of The Cariboo's Learning Network (British Columbia), and Washtenaw Community College (Michigan). Each chapter provides an overview of the program, including a description of the students served, the program organization, instructional strategies utilized, technologies employed, student services provided, and student grading and program evaluation methods, as well as a discussion of the unique or exemplary practices involved. An annotated list of DE programs at 38 colleges is appended. Contains 15 references. (HAA)

At the Homestead Campus of Miami-Dade Community College (MDCC), in Florida, one barrier to fully integrating technology in the classroom was identified as the low-level of student computer skills. To address the problem, the campus undertook a project to establish a technology basic skills workshop for all students. In informal interviews to determine basic technology skills, faculty indicated that all students should be able to use a word processor, use spell- and grammar-checkers for proofing and editing, use computer tutorials, and use CD-ROMs for research. The technology workshop was then designed as a one-credit, facilitated course consisting of the following five self-paced modules: (1) an introduction to the facilitated learning format; (2) an introduction to technology lab skills; (3) a guide to Microsoft Works software; (4) an introduction to MDCC's computerized advisement and registration system; and (5) a guide to using the campus library. In implementing the program, the campus had to overcome issues related to logistics, such as tracking students, and to making the workshop economically productive. Additional benefits of the workshop were that students gained skills necessary to function on campus and in the workplace, faculty have begun to assign more computer-based work, and both the campus and the technologies have been made more accessible to the community. (KP)

This resource guide provides information on distance education options and offers a directory of programs at various institutions in the United States and Canada. Following a brief preface on the development of the guide, "New Models for Higher Education in Changing Times," by Don Doucette, discusses the nature of distance learning, the inevitability of this type of instructional delivery, the infrastructure provided by information technology, and changes required in the current model of higher education. Next, an introductory essay outlines the basics of distance education, explaining its role as an alternative to traditional classroom instruction and outlining the following forms of distance education: (1) print, including correspondence courses with textbooks, study guides, and assignments mailed to students; (2) audio, featuring radio, audiocassette, and telephone-based instruction; (3) video, including television, videocassette, and videodisc courses; and (4) computer-based approaches, featuring telecommunications, electronic mail, electronic conferencing, electronic bulletin boards, and telecommunications networks. Results of a survey of 1,000 community colleges are also included, indicating the top delivery media used and the most and least frequently offered courses. The bulk of the guide consists of a list of community college distance education programs by state or province. Each entry provides the college's name and address, the media used and disciplines taught in distance education, and the name of a contact person. Explanatory notes highlighting special features are included with some entries. (MAB)

The Homestead Campus of Miami-Dade Community College, in Florida, serves a sparsely populated area with a culturally diverse population including migrant farm workers, prison inmates, and U.S. Air Force personnel. To increase access to college services, the campus focused on implementing a computer-based distance education program as its primary technology initiative. The campus undertook an analysis of the costs and benefits associated with implementing the program to determine its viability as an academic alternative. Benefits determined for the distance education program included an increased market, allowing the campus to expand enrollment; improved productivity, by embracing the concept of student-centered facilitated learning; and decreased facilities investment, since distance programs do not require physical space. Considerable costs were also determined, however, including costs for equipment acquisition or adaptation to expand existing computer networks and electronic mail systems; for personnel in the areas of technical support, technology training, instructional design, and program support; and for curriculum development to adapt courses to distance delivery. While start-up costs of distance education can be high, the return on investment can be significant, providing a wide array of institutional benefits. (KP)

Although technology is potentially a powerful tool for improving the teaching-learning process, studies have shown that it is widely underutilized, with many faculty lacking expertise in and awareness of available technologies. Closing this gap between the availability of technology and faculty ownership of it is the function of an instructional technology transfer center. Strategies used by 11 representative centers to enhance and improve teaching and learning at community colleges include efforts targeted at faculty as a group, at individual teachers, and at the campus as a whole. Group strategies include conducting needs assessments to prioritize training interests, offering day-long workshops to train faculty to use software packages or mini-workshops of less than 2 hours to cover such specific topics as how to use a laser disc or test-generating programs, arranging vendor demonstrations of products, and establishing special projects and faculty teams. Strategies focusing on individuals include providing materials for independent learning; establishing curriculum application projects, offering release time, professional development awards, and monetary compensation; designating proficient faculty as mentors; and loaning hardware and/or software during academic breaks. Finally, campus-wide initiatives include newsletters to keep faculty informed of center activities, open houses and orientation sessions for new employees, computer fairs, involving faculty on center steering committees, and acting as a clearinghouse of information for vendors. Contains seven references. (KP)

Teaching methods seem to work with some students and fail with others, because every learner is unique. Teachers must be able to identify learning styles and find ways for instruction to match those styles. This paper moves from an overview of different learning styles to a review of the use of computer-assisted instruction. A recent development in technology, known as hypermedia, is described as a tool for faculty training and managing curricula. Hypermedia is an expansion of the hypertext concept that allows for the inclusion of media links as well as data links. Information in various types of electronic media can be accessed by activating the link. The user can explore information in a direction and speed suited to his or her learning style, as well as through the sensory stimuli that best enhances learning. The HyperCard software package provides tools to permit the educator-author the freedom and flexibility to address the needs of any learner. While development efforts may require more time when using technology to provide alternative learning paths, only minor revisions should be needed once the original work is done. (Contains 10 references.) (SLD)

The analysis determined the extent to which 1972 occupational projections have changed from 1966 for high school sophomores, investigating patterns as they related to race, sex, and residence. The samples consisted of all sophomores in the high schools of 3 nonmetropolitan low income countries in East Texas (Burleson, Leon, and San Jacinto). Chi square tests of difference estimated statistically significant changes. The 20 tables give percentage results, both by race and sex, and by race and residence, for changes in occupational aspirations and aspiration levels, occupational expectations and expectation levels, intensity of occupational aspiration and level of intensity, certainty of occupational expectations, and residence and occupational projections. The investigation concluded that, for most categories, there was considerable stability between 1966 and 1972. (KM)

The changes in residence projections (statements of behavior, attitudes, or beliefs about future residence status) of rural East Texas high school sophomores between 1966 and 1972 were examined. Focus was on general trends over the time span as well as those specific to race, sex, and respondent's residence. The analysis involved an age-education-level cohort in order to estimate changes due to factors originating in the individual's environment rather than those associated with individual development over a given time span. Types of residence projections examined were: aspirations, expectations, anticipatory deflection, aspiration intensity, and expectation certainty. Respondents (473 sophomores in 1966 and 461 in 1972) were from 3 low income rural counties in East Texas (Burleson, Leon, and San Jacinto) which had high proportions of blacks. The same group-administered questionnaire and administration procedures were used both years. Some findings were: (1) the youth had shifted to some extent away from urban and toward rural residences in their aspirations and expectations; (2) the shift was greater for white than for black youth and greater for males than females; and (3) black females were less certain in 1972 than 1966 of attaining their residence expectation. (NQ)

The status projections of high school sophomores residing in rural East Texas were surveyed in spring of 1972. Focus was on their aspirations and expectations for status attainment in occupation, education, income, and type of place of residence. In each area, four elements of status projections were examined: aspiration level, expectation level, certainty of expectation, and intensity of aspiration. A small-scale investigation of the observations made in the initial study was later conducted with 11 respondents from a school where the interview situation had been far from ideal. This study evaluated observation reliability, in terms of consistency of responses, through a "test-retest" procedure over a 2-week time lapse. Data were collected via the same group-administered questionnaire used in the initial study. Among the findings were: (1) variation in response, in terms of initial coded measurements, was relatively high across all status areas, except for status object of aspirations; and (2) generally, indicators for all elements of educational status projections were more stable than those of other status areas. The general intent of this study was to produce some reasonable hypotheses for additional and broader research since the limited sample does not allow for generalizations. (NQ)