arXiv subjects

I. Y. Tyukin

Publications and source records attributed to I. Y. Tyukin.

7 recordsLinked to original sources

arXiv2019

How deep should be the depth of convolutional neural networks: a backyard dog case study

The work concerns the problem of reducing a pre-trained deep neuronal network to a smaller network, with just few layers, whilst retaining the network's functionality on a given task The proposed approach is motivated by the observation that the aim to deliver the highest accuracy possible in the broadest range of operational conditions, which many deep neural networks models strive to achieve, may not necessarily be always needed, desired, or even achievable due to the lack of data or technical constraints. In relation to the face recognition problem, we formulated an example of such a usecase, the `backyard dog' problem. The `backyard dog', implemented by a lean network, should correctly identify members from a limited group of individuals, a `family', and should distinguish between them. At the same time, the network must produce an alarm to an image of an individual who is not in a member of the family. To produce such a network, we propose a shallowing algorithm. The algorithm takes an existing deep learning model on its input and outputs a shallowed version of it. The algorithm is non-iterative and is based on the Advanced Supervised Principal Component Analysis. Performance of the algorithm is assessed in exhaustive numerical experiments. In the above usecase, the `backyard dog' problem, the method is capable of drastically reducing the depth of deep learning neural networks, albeit at the cost of mild performance deterioration. We developed a simple non-iterative method for shallowing down pre-trained deep networks. The method is generic in the sense that it applies to a broad class of feed-forward networks, and is based on the Advanced Supervise Principal Component Analysis. The method enables generation of families of smaller-size shallower specialized networks tuned for specific operational conditions and tasks from a single larger and more universal legacy network.

cs.NE↗

arXiv2018

Simple model of complex dynamics of activity patterns in developing networks of neuronal cultures

Living neuronal networks in dissociated neuronal cultures are widely known for their ability to generate highly robust spatiotemporal activity patterns in various experimental conditions. These include neuronal avalanches satisfying the power scaling law and thereby exemplifying self-organized criticality in living systems. A crucial question is how these patterns can be explained and modeled in a way that is biologically meaningful, mathematically tractable and yet broad enough to account for neuronal heterogeneity and complexity. Here we propose a simple model which may offer an answer to this question. Our derivations are based on just few phenomenological observations concerning input-output behavior of an isolated neuron. A distinctive feature of the model is that at the simplest level of description it comprises of only two variables, a network activity variable and an exogenous variable corresponding to energy needed to sustain the activity and modulate the efficacy of signal transmission. Strikingly, this simple model is already capable of explaining emergence of network spikes and bursts in developing neuronal cultures. The model behavior and predictions are supported by empirical observations and published experimental evidence on cultured neurons behavior exposed to oxygen and energy deprivation. At the larger, network scale, introduction of the energy-dependent regulatory mechanism enables the network to balance on the edge of the network percolation transition. Network activity in this state shows population bursts satisfying the scaling avalanche conditions. This network state is self-sustainable and represents a balance between global network-wide processes and spontaneous activity of individual elements.

nlin.AO↗

arXiv2018

Correction of AI systems by linear discriminants: Probabilistic foundations

Artificial Intelligence (AI) systems sometimes make errors and will make errors in the future, from time to time. These errors are usually unexpected, and can lead to dramatic consequences. Intensive development of AI and its practical applications makes the problem of errors more important. Total re-engineering of the systems can create new errors and is not always possible due to the resources involved. The important challenge is to develop fast methods to correct errors without damaging existing skills. We formulated the technical requirements to the 'ideal' correctors. Such correctors include binary classifiers, which separate the situations with high risk of errors from the situations where the AI systems work properly. Surprisingly, for essentially high-dimensional data such methods are possible: simple linear Fisher discriminant can separate the situations with errors from correctly solved tasks even for exponentially large samples. The paper presents the probabilistic basis for fast non-destructive correction of AI systems. A series of new stochastic separation theorems is proven. These theorems provide new instruments for fast non-iterative correction of errors of legacy AI systems. The new approaches become efficient in high-dimensions, for correction of high-dimensional systems in high-dimensional world (i.e. for processing of essentially high-dimensional data by large systems).

cs.LG↗

arXiv2018

The unreasonable effectiveness of small neural ensembles in high-dimensional brain

Despite the widely-spread consensus on the brain complexity, sprouts of the single neuron revolution emerged in neuroscience in the 1970s. They brought many unexpected discoveries, including grandmother or concept cells and sparse coding of information in the brain. In machine learning for a long time, the famous curse of dimensionality seemed to be an unsolvable problem. Nevertheless, the idea of the blessing of dimensionality becomes gradually more and more popular. Ensembles of non-interacting or weakly interacting simple units prove to be an effective tool for solving essentially multidimensional problems. This approach is especially useful for one-shot (non-iterative) correction of errors in large legacy artificial intelligence systems. These simplicity revolutions in the era of complexity have deep fundamental reasons grounded in geometry of multidimensional data spaces. To explore and understand these reasons we revisit the background ideas of statistical physics. In the course of the 20th century they were developed into the concentration of measure theory. New stochastic separation theorems reveal the fine structure of the data clouds. We review and analyse biological, physical, and mathematical problems at the core of the fundamental question: how can high-dimensional brain organise reliable and fast learning in high-dimensional world of data by simple tools? Two critical applications are reviewed to exemplify the approach: one-shot correction of errors in intellectual systems and emergence of static and associative memories in ensembles of single neurons.

cs.AI↗

arXiv2018

Blessing of dimensionality: mathematical foundations of the statistical physics of data

The concentration of measure phenomena were discovered as the mathematical background of statistical mechanics at the end of the XIX - beginning of the XX century and were then explored in mathematics of the XX-XXI centuries. At the beginning of the XXI century, it became clear that the proper utilisation of these phenomena in machine learning might transform the curse of dimensionality into the blessing of dimensionality. This paper summarises recently discovered phenomena of measure concentration which drastically simplify some machine learning problems in high dimension, and allow us to correct legacy artificial intelligence systems. The classical concentration of measure theorems state that i.i.d. random points are concentrated in a thin layer near a surface (a sphere or equators of a sphere, an average or median level set of energy or another Lipschitz function, etc.). The new stochastic separation theorems describe the thin structure of these thin layers: the random points are not only concentrated in a thin layer but are all linearly separable from the rest of the set, even for exponentially large random sets. The linear functionals for separation of points can be selected in the form of the linear Fisher's discriminant. All artificial intelligence systems make errors. Non-destructive correction requires separation of the situations (samples) with errors from the samples corresponding to correct behaviour by a simple and robust classifier. The stochastic separation theorems provide us by such classifiers and a non-iterative (one-shot) procedure for learning.

cs.LG↗

arXiv2017

Stochastic Separation Theorems

The problem of non-iterative one-shot and non-destructive correction of unavoidable mistakes arises in all Artificial Intelligence applications in the real world. Its solution requires robust separation of samples with errors from samples where the system works properly. We demonstrate that in (moderately) high dimension this separation could be achieved with probability close to one by linear discriminants. Surprisingly, separation of a new image from a very large set of known images is almost always possible even in moderately high dimensions by linear functionals, and coefficients of these functionals can be found explicitly. Based on fundamental properties of measure concentration, we show that for $M 1-\vartheta$, where $1>\vartheta>0$ is a given small constant. Exact values of $a,b>0$ depend on the probability distribution that determines how the random $M$-element sets are drawn, and on the constant $\vartheta$. These {\em stochastic separation theorems} provide a new instrument for the development, analysis, and assessment of machine learning methods and algorithms in high dimension. Theoretical statements are illustrated with numerical examples.

cs.LG↗

arXiv2003

Adaptive Algorithms in Finite Forms

We propose a new method to design adaptation algorithms that guarantee a certain prescribed level of performance and are applicable to systems with nonconvex parameterization. The main idea behind the method is, given the desired performance characteristics and a class of nonlinearly parameterized systems, first to augment the tuning error function and design the adaptation scheme in the form of ordinary differential equations. The resulting augmentation is allowed to depend on state derivatives. To deal with these, we suggest the realization of the adaptation scheme in an algebraic-integral form. Because of the explicit dependance on the state of the original system such adaptation schemes are referred to as adaptive algorithms in it finite form instead of differential ones. Sufficient conditions for the existence of finite form realizations are proposed. These conditions lead to the necessity to find a solution of a system of partial differential equations, which in general is not an easy task. In order to resolve this problem we suggest to embed the original system dynamics into one of a higher order, thus replacing it by a simple integration of a function with respect to a single scalar argument. Several full-state feedback finite form realizations are presented and illustrated with examples.

math.OC↗