Basabi Bhaumik

Optimization with neural networks: a recipe for improving convergence and solution quality

Artificial Neural Networks, particularly the Hopfield Network have been applied to the solution of a variety of tasks formulated as optimization problems. However, the network often converges to invalid solutions, which have been attributed to an improper choice of parameters and energy functions. In this letter, we propose a fundamental change of viewpoint. We assert that the problem is not due to the bad choice of parameters or the form of the energy function chosen. Instead, we show that the Hopfield Net essentially performs only one iteration of a Sequential Unconstrained Minimization Technique (SUMT). Thus, it is not surprising that unsatisfactory results are obtained. We present results on an SUMT-based formulation for the Travelling Salesman Problem, where we consistently obtained valid tours. We also show how shorter tours can be systematically obtained.

A low power 256 KB SRAM design

In this paper a low power SRAM design is presented. Existing SRAM architectures used in SGS Thomson were studied to explore the possibilities in bringing down power dissipation in various blocks. A Divided word line (DWL) scheme was implemented. Particular emphasis was put to reduce power consumption in decoders. A new critical path model was introduced for schematic simulation. This lowered the simulation time considerably. Simulation results confirmed the effectiveness of our approach.

The Balance Between Excitation and Inhibition Not Only Leads to Variable Discharge of Cortical Neurons but Also to Contrast Invariant Orientation Tuning

The orientation tuning width of the spike response of neuron in layer V1 of primary visual cortex does not change with contrast of input signal. It is also known that cortical neurons exhibit tremendous irregularity in their discharge pattern which is conserved over large regions of cerebral cortex. To produce this irregularity in responses the neurons must receive balanced excitation and inhibition. By a modeling study we show that if this balance is maintained for all levels of contrast it results in variable discharge patterns of cortical neurons at all contrast and also in contrast invariant orientation tuning. Further, this study supports the role of inhibition in shaping the responses of cortical neurons and we also obtain changes in circular variance with changing contrast, similar to what is observed experimentally.

Orientation map: a reaction-diffusion based model

We show that models for orientation map formation, which invoke convolution in the dynamic equation, are in fact invoking non-local reaction diffusion. In the early stages of development it is unlikely that non-local interactions take place. We present a nonlinear reaction diffusion model based on near neighbour interaction for formation of orientation map. Besides capturing all the essential topological qualities of experimental orientation, our model also deals with the interspecies variations in the orientation map.

Development of feedforward receptive field structure of a simple cell and its contribution to the orientation selectivity: a modeling study.

Recent experimental studies of hetero-synaptic interactions in various systems have shown the role of signaling in the plasticity, challenging the conventional understanding of Hebbs rule. It has also been found that activity plays a major role in plasticity, with neurotrophins acting as molecular signals translating activity into structural changes. Furthermore, role of synaptic efficacy in biasing the outcome of competition has also been revealed recently. Motivated by these experimental findings we present a model for the development of simple cell receptive field structure based on the competitive hetero-synaptic interactions for neurotrophins combined with cooperative hetero-synaptic interactions in the spatial domain. We find that with proper balance in competition and cooperation, the inputs from two populations (ON/OFF) of LGN cells segregate starting from the homogeneous state. We obtain segregated ON and OFF regions in simple cell receptive field. Our modeling study supports the experimental findings, suggesting the role of synaptic efficacy and the role of spatial signaling. We find that using this model we obtain simple cell RF, even for positively correlated activity of ON/OFF cells. We also compare different mechanism of finding the response of cortical cell and study their possible role in the sharpening of orientation selectivity. We find that degree of selectivity improvement in individual cells varies from case to case depending upon the structure of RF field and type of sharpening mechanism.

Development of a Simple Cell Receptive Field Structure: A Model Based on Hetero-synaptic Interactions

Recent experimental studies of hetero-synaptic interactions in various systems have shown the role of spatial signaling in plasticity, challenging the conventional understanding of Hebb’s rule. It has also been found that activity plays a major role in plasticity, with neurotrophins acting as molecular signals translating activity into structural changes. Furthermore, role of synaptic efficacy in biasing the outcome of competition has also been revealed recently. Motivated by these experimental findings we present a model for the development of a simple cell receptive field structure based on competitive and cooperative hetero-synaptic interaction in the spatial domain. We find that with proper balance of competition and cooperation, the inputs from the two populations (ON/OFF) of LGN cells segregate starting from the homogeneous state. We obtain segregated ON and OFF regions in simple cell receptive field.

Diffusive Hebbian model for orientation map formation

The distribution of orientation selective cells over the cortical surface is captured in the orientation map. In the existing models on orientation map formation, the lateral connectivity in the cortex is modelled by Mexican hat intra-cortical interaction. The Mexican hat intra-cortical interaction requires long range interaction to be present during early period of development. We show that near neighbour diffusive Hebbian learning mechanism along with self-inhibition in the cells is sufficient for the formation of an orientation map. Long range intra-cortical connectivity is not critical for formation of orientation maps.

Comments on "A naive approach to global optimization" by L.O. Chua

For original paper see ibid., vol.37, p.966-9 (1990). The commenters point out certain practical problems with a recently published approach to global optimization, and also describe briefly an algorithm which does not suffer from these problems. They claim that their method can be implemented in circuit form and using neural networks. >

Ratio of average inhibitory to excitatory conductance modulates the response of simple cell

Recent experimental study reports existence of complex type of interneurons in the primary visual cortex. The response of these inhibitory cells depends mainly upon feed-forward LGN inputs. The goal of this study is to determine the role of these cells in modulating the response of simple cells. Here we demonstrate that if the inhibitory contribution due to these cells balances the feed-forward excitatory inputs the spike response of cortical cell becomes sharply tuned. Using a single cell integrate and fire neuron model we show that the ratio of average inhibitory to excitatory conductance controls the balance between excitation and inhibition. We find that many different values of ratio can result in balanced condition. However, the response of the cell is not sharply tuned for each of these ratios. In this study we explicitly determine the best value of ratio needed to make the response of the cell sharply tuned.

A Microscopic Framework For Distributed Object-Recognition & Pose-Estimation

Effective self-organization schemes lead to the creation of autonomous and reliable robot teams that can outperform a single, sophisticated robot on several tasks. We present here a novel, vision-based microscopic framework for active and distributed object-recognition and pose-estimation using a team of robots of simple construction. The team performs the task of locating a given object(s) in an unknown territory, recognizing it with sufficient confidence and estimating its pose. The larger goal is to experiment with probabilistic frameworks and graph-theoretic methods in the design of robot teams to achieve autonomous self-organization independent of the task at hand. We have chosen 3D object recognition as a first problem area to evaluate the effectiveness of our system design. The system comprises a probabilistic framework for the successful detection of the object in a coordinated manner and adaptive measures in case of machinery failures or presence of obstacles. A pose estimation method for the detected object and graph theoretic solutions for optimal field coverage by the robots are also presented. Each robot is provided with a part-based, spatial model of the object. The object to be recognized is taken to be much bigger than the robots and need not fit completely into the field of view of the robot cameras. We assume no knowledge of the internal parameters of the robot cameras and perform no camera calibration procedures. Initial simulation results corroborate our system design and field coverage methods