Semi-automatic detection of honeybee brood hygiene-an example of artificial learning to facilitate ethological studies on social insects.

Machine-learning techniques are shifting the boundaries of feasibility in many fields of ethological research. Here, we describe an application of machine learning to the detection/measurement of hygienic behaviour, an important breeding trait in the honey bee (Apis mellifera). Hygienic worker bees are able to detect and destroy diseased brood, thereby reducing the reproduction of economically important pathogens and parasites such as the Varroa mite (Varroa destructor). Video observation of this behaviour on infested combs has many advantages over other methods of measurement, but analysing the recorded material is extremely time-consuming. We approached this problem by combining automatic tracking of bees in the video recordings, extracting relevant features, and training a multi-layer discriminator on positive and negative examples of the behaviour of interest. Including expert knowledge into the design of the features lead to an efficient model for identifying the uninteresting parts of the video which can be safely skipped. This algorithm was then used to semiautomatically identify individual worker bees involved in the behaviour. Application of the machine-learning method allowed to save 70% of the time required for manual analysis, and substantially increased the number of cell openings correctly identified. It thereby turns video-observation of individual cell opening events into an economically competitive method for selecting potentially resistant bees. This method presents an example of how machine learning can be used to boost ethological research, and how it can generate new knowledge by explaining the learned decision rule in form of meaningful parameters.

Approximate Bayes learning of stochastic differential equations.

We introduce a nonparametric approach for estimating drift and diffusion functions in systems of stochastic differential equations from observations of the state vector. Gaussian processes are used as flexible models for these functions, and estimates are calculated directly from dense data sets using Gaussian process regression. We develop an approximate expectation maximization algorithm to deal with the unobserved, latent dynamics between sparse observations. The posterior over states is approximated by a piecewise linearized process of the Ornstein-Uhlenbeck type and the maximum a posteriori estimation of the drift is facilitated by a sparse Gaussian process approximation.

Successful attack on permutation-parity-machine-based neural cryptography.

An algorithm is presented which implements a probabilistic attack on the key-exchange protocol based on permutation parity machines. Instead of imitating the synchronization of the communicating partners, the strategy consists of a Monte Carlo method to sample the space of possible weights during inner rounds and an analytic approach to convey the extracted information from one outer round to the next one. The results show that the protocol under attack fails to synchronize faster than an eavesdropper using this algorithm.

Switching regulatory models of cellular stress response.

MOTIVATION: Stress response in cells is often mediated by quick activation of transcription factors (TFs). Given the difficulty in experimentally assaying TF activities, several statistical approaches have been proposed to infer them from microarray time courses. However, these approaches often rely on prior assumptions which rule out the rapid responses observed during stress response. RESULTS: We present a novel statistical model to infer how TFs mediate stress response in cells. The model is based on the assumption that sensory TFs quickly transit between active and inactive states. We therefore model mRNA production using a bistable dynamical systems whose behaviour is described by a system of differential equations driven by a latent stochastic process. We assume the stochastic process to be a two-state continuous time jump process, and devise both an exact solution for the inference problem as well as an efficient approximate algorithm. We evaluate the method on both simulated data and real data describing Escherichia coli's response to sudden oxygen starvation. This highlights both the accuracy of the proposed method and its potential for generating novel hypotheses and testable predictions. AVAILABILITY: MATLAB and C++ code used in the article can be downloaded from http://www.dcs.shef.ac.uk/~guido/.

Efficient statistical inference for stochastic reaction processes.

We address the problem of estimating unknown model parameters and state variables in stochastic reaction processes when only sparse and noisy measurements are available. Using an asymptotic system size expansion for the backward equation, we derive an efficient approximation for this problem. We demonstrate the validity of our approach on model systems and generalize our method to the case when some state variables are not observed.

Dynamics of neural cryptography.

Synchronization of neural networks has been used for public channel protocols in cryptography. In the case of tree parity machines the dynamics of both bidirectional synchronization and unidirectional learning is driven by attractive and repulsive stochastic forces. Thus it can be described well by a random walk model for the overlap between participating neural networks. For that purpose transition probabilities and scaling laws for the step sizes are derived analytically. Both these calculations as well as numerical simulations show that bidirectional interaction leads to full synchronization on average. In contrast, successful learning is only possible by means of fluctuations. Consequently, synchronization is much faster than learning, which is essential for the security of the neural key-exchange protocol. However, this qualitative difference between bidirectional and unidirectional interaction vanishes if tree parity machines with more than three hidden units are used, so that those neural networks are not suitable for neural cryptography. In addition, the effective number of keys which can be generated by the neural key-exchange protocol is calculated using the entropy of the weight distribution. As this quantity increases exponentially with the system size, brute-force attacks on neural cryptography can easily be made unfeasible.

Genetic attack on neural cryptography.

Different scaling properties for the complexity of bidirectional synchronization and unidirectional learning are essential for the security of neural cryptography. Incrementing the synaptic depth of the networks increases the synchronization time only polynomially, but the success of the geometric attack is reduced exponentially and it clearly fails in the limit of infinite synaptic depth. This method is improved by adding a genetic algorithm, which selects the fittest neural networks. The probability of a successful genetic attack is calculated for different model parameters using numerical simulations. The results show that scaling laws observed in the case of other attacks hold for the improved algorithm, too. The number of networks needed for an effective attack grows exponentially with increasing synaptic depth. In addition, finite-size effects caused by Hebbian and anti-Hebbian learning are analyzed. These learning rules converge to the random walk rule if the synaptic depth is small compared to the square root of the system size.

Neural cryptography with feedback.

Neural cryptography is based on a competition between attractive and repulsive stochastic forces. A feedback mechanism is added to neural cryptography which increases the repulsive forces. Using numerical simulations and an analytic approach, the probability of a successful attack is calculated for different model parameters. Scaling laws are derived which show that feedback improves the security of the system. In addition, a network with feedback generates a pseudorandom bit sequence which can be used to encrypt and decrypt a secret message.