Larry Jackel

North C Technologies

title>ABSTRACT A promising approach to autonomous driving is machine learning. In machine learning systems, training datasets are created that capture the sensory input to a vehicle as well as the desired response. One disadvantage of using a learned navigation system is that the learning process itself may require both a huge number of training e...

In this paper, we present AmbiSense, an acoustic field based sensing system that performs proximity detection and bearing estimation for safer physical human-robot interactions. A single low cost piezoelectric transducer is used to setup this novel acoustic sensing modality to create a blindspot-free sound field engulfing a robot arm. Two detection...

Perceiving obstacles and avoiding collisions is fundamental to the safe operation of a robot system, particularly when the robot must operate in highly dynamic human environments. Proximity detection using on-robot sensors can be used to avoid or mitigate impending collisions. However, existing proximity sensing methods are orientation and placemen...

Perceiving obstacles and avoiding collisions is fundamental to the safe operation of a robot system, particularly when the robot must operate in highly dynamic human environments. Proximity detection using on-robot sensors can be used to avoid or mitigate impending collisions. However, existing proximity sensing methods are orientation and placemen...

Collision detection is critical for safe robot operation in the presence of humans. Acoustic information originating from collisions between robots and objects provides opportunities for fast collision detection and localization; however, audio information from microphones on robot manipulators needs to be robustly differentiated from motors and ex...

The U.S. Defense Advanced Research Projects Agency (DARPA) has played a remarkable role in the creation new transformative technologies, revolutionizing defense with drones and precision-guided munitions, and transforming civilian life with portable GPS receivers, voice-recognition software, self-driving cars, unmanned aerial vehicles, and, most fa...

The DARPA Robotics Challenge (DRC) program conducted a series of prize-based competition events to develop and demonstrate technology for disaster response. This chapter provides the official and definitive account of DRC Finals as the culmination of the DRC program. The chapter details the eight tasks (Drive, Egress, Door, Valve, Wall, Surprise (P...

As part of a complete software stack for autonomous driving, NVIDIA has created a neural-network-based system, known as PilotNet, which outputs steering angles given images of the road ahead. PilotNet is trained using road images paired with the steering angles generated by a human driving a data-collection car. It derives the necessary domain know...

Seamless Electronics for Automotive Services. Going forward from the last ELIV „Electronics in Vehicles“ in 2015 – the most significant Congress in Automotive Electronics has now seen a substantial upgrade. In line with the feedback given by participants, speakers and journalists we have added new elements and contents to the event, which is benefi...

This paper proposes a new method, that we call VisualBackProp, for visualizing which sets of pixels of the input image contribute most to the predictions made by the convolutional neural network (CNN). The method heavily hinges on exploring the intuition that the feature maps contain less and less irrelevant information to the prediction decision w...

The DARPA Robotics Challenge (DRC) program conducted a series of prize-based competition events to develop and demonstrate technology for disaster response. This article provides the official and definitive account of DRC Finals as the culmination of the DRC program. The article details the eight tasks (Drive, Egress, Door, Valve, Wall, Surprise [P...

A promising approach to autonomous driving is machine learning. In such systems, training datasets are created that capture the sensory input to a vehicle as well as the desired response. A disadvantage of using a learned navigation system is that the learning process itself may require a huge number of training examples and a large amount of compu...

We trained a convolutional neural network (CNN) to map raw pixels from a single front-facing camera directly to steering commands. This end-to-end approach proved surprisingly powerful. With minimum training data from humans the system learns to drive in traffic on local roads with or without lane markings and on highways. It also operates in areas...

We are developing a complete, self-contained autonomous navigation system for mobile robots that learns quickly, uses commodity components, and has the added benefit of emitting no radiation signature. It builds on the au­tonomous navigation technology developed by Net-Scale and New York University during the Defense Advanced Research Projects Agen...

Two Defense Advanced Research Projects Agency (DARPA) robotics programs have been formulated for enhancing the development of autonomous locomotion and navigation. The projects have developed simple experimental methodology for evaluating growth and promoting cross-team cooperation. The programs have accomplished the impossible task of autonomous l...

Due to the technology available, most previous work in teleoperated robotics used relatively low-resolution video links and provided limited perceptual feedback to the teleoperator. In most cases, these projects reported only limited teleoperator success compared to vehicles with human drivers on-board. We set out to build a high-fidelity teleopera...

This article describes the conduct of six evaluation experiments for the Perception for Off-Road Robotics program. Key distinctions of the testing methodology include conduct of the experiments by a group independent from the developers, unrehearsed experiments that provide little advance knowledge of the test courses, and blind experiments that do...

The DARPA Learning Applied to Ground Vehicles (LAGR) program is accelerating progress in autonomous, perception-based, off-road navigation in unmanned ground vehicles (UGVs) by incorporating learned behaviors. In addition, the program is using passive optical systems to accomplish long-range scene analysis. By combining long-range perception with l...

Lithography ('stone writing') is used to describe the general writing and printing processes by which desired patterns are created in or transferred to solid substrates. In recent years lithography techniques have been developed for use in the nanometer domain. The nanometer domain covers sizes bigger than several atoms but smaller than the wavelen...

We have developeda neural-network architecture for recognizing

This paper compares the performance of several classifier algorithms on a standard database of handwritten digits. We consider not only raw accuracy, but also training time, recognition time, and memory requirements. When available, we report measurements of the fraction of patterns that must be rejected so that the remaining patterns have misclass...

This paper compares the performance of several classifier algorithms on a standard database of handwritten digits. We consider not only raw accuracy, but also training time, recognition time, and memory requirements. When available, we report measurements of the fraction of patterns that must be rejected so that the remaining patterns have misclass...

Random errors and insufficiencies in databases limit the performance of any classifier trained from and applied to the database. In this paper we propose a method to estimate the limiting performance of classifiers imposed by the database. We demonstrate this technique on the task of predicting failure in telecommunication paths.

This paper compares the performance of classifier algorithms on a standard database of handwritten digits. We consider not raw accuracy, but rejection, training time, recognition time, and memory requirements. "Comparison of Leaning for Handwritten Digit Recognition", International Conference on Neural F. and P. Cie Publishers, 1995 Y. Le L. Bottou...

This paper compare the performance of classifier algorithms on a stan- dard database of handwritten digits. We consider not only raw accuracy, but also training time, recognition time, and memory requirements. When available, we report measurements of the fraction of patterns that must be rejected so that the remaining patterns have misclassificati...

This paper compares the performance of several classi#er algorithms on a standard database of handwritten digits. We consider not only raw accuracy, but also rejection, training time, recognition time, and memory requirements. 1 COMPARISON OF LEARNING ALGORITHMS FOR HANDWRITTEN DIGIT RECOGNITION Y. LeCun, L. Jackel, L. Bottou, A. Brunot, C. Cortes,...

is network has a 20#20 pixel input #eld and ten outputs for the classi#cation result. The #rst three layers, which contain 97 # of the connections and thus constitute the main computational load, have been implemented on a single ANNA chip. The preprocessing #size normalization and deskewing# as well as the last two layers #which require higher pre...

A board is described that contains the ANNA neural-network chip, and a DSP32C digital signal processor. The ANNA #Analog Neural Network Arithmetic unit# chip performs mixed analog#digital processing. The combination of ANNA with the DSP allows high-speed, end-toend execution of numerous signal-processing applications, including the preprocessing, t...

We present an application of back-propagation networks to handwritten digit recognition. Minimal preprocessing of the data was required, but architecture of the network was highly constrained and specifically designed for the task. The input of the network consists of normalized images of isolated digits. The method has 1% error rate and about a 9%...

This paper compares the performance of classifier algorithmson a standard database of handwritten digits. We consider not rawaccuracy, but rejection, training time, recognition time, and memoryrequirements."Comparison of Leaning for Handwritten Digit Recognition", International Conference onNeural F. and P. Cie Publishers, 1995Y. Le L. Bottou, C. J...

Character Recognition has served as one of the principal proving grounds for neural-net methods and has emerged as one of the most successful applications of this technology. This chapter outlines optical character recognition document analysis systems developed at AT&T Bell Labs that combine the strengths of machine-learning algorithms with high-s...

In an optical character recognition problem, we compare (as a function of training set size) the performance of three neural network based ensemble methods (two versions of boosting and a committee of neural networks trained independently) to that of a single network. In boosting, the number of patterns actually used for training is a subset of all...

This paper compares the performance of several classifier algorithms on a standard database of handwritten digits. We consider not only raw accuracy, but also training time, recognition time, and memory requirements. When available, we report measurements of the fraction of patterns that must be rejected so that the remaining patterns have misclass...

We compare the performance of three types of neural network-based ensemble techniques to that of a single neural network. The ensemble algorithms are two versions of boosting and committees of neural networks trained independently. For each of the four algorithms, we experimentally determine the test and training error curves in an optical characte...

Random errors and insufficiencies in databases limit the performance of any classifier trained from and applied to the database. In this pa- per we propose a method to estimate the lim- iting performance of classifiers imposed by the database. We demonstrate this technique on the task of predicting failure in telecommunication paths.

We present a survey of recent electronic implementations of neural nets in the US and Canada with an emphasis on integrated circuits. Well over 50 different circuits were built during the last two years, representing a remarkable variety of designs. They range from digital emulators to fully analog CMOS networks operating in the subthreshold region...

The authors present a new feature extraction method together with neural network recognition for online Chinese characters. A Chinese character can be represented by a three-dimensional 12 × 12 × 4 array of numbers. Multiple conventional neural networks are used for online small vocabulary Chinese character recognition based on this feature extract...

Two of our analog neural net chips have been integrated into board systems and are being used now in a variety of image recognition applications. One of the two circuits, the NET32K chip, has connections with a low resolution of between one and four bits. With this chip one can scan up to 32 kernels of a size of 16 X 16 pixels over an image. It is...

A neural network algorithm-based system that reads handwritten ZIP codes appearing on real US mail is described. The system uses a recognition-based segmenter, that is a hybrid of connected-components analysis (CCA), vertical cuts, and a neural network recognizer. Connected components that are single digits are handled by CCA. CCs that are combined...

A neural network with 136000 connections for recognition of handwritten digits has been implemented using a mixed analog/digital neural network chip. The neural network chip is capable of processing 1000 characters/s. The recognition system has essentially the same rate (5%) as a simulation of the network with 32-b floating-point precision

A special-purpose chip, optimized for computational needs of neural networks and performing over 2000 multiplications and additions simultaneously, is described. Its data path is particularly suitable for the convolutional architectures typical in pattern classification networks but can also be configured for fully connected or feedback topologies....

We describe two neural-net approaches to digit recognition. One method uses a neural-network chip to perform line thinning and local feature extraction. This preprocessing stage was designed by hand and did not involve any learning. However, automatic learning was used in the final classification step. The chip can process about 100 characters/sec,...

The authors describe a method which combines dynamic programming and a neural network recognizer for segmenting and recognizing character strings. The method selects the optimal consistent combination of cuts from a set of candidate cuts generated using heuristics. The optimal segmentation is found by representing the image, the candidate segments,...

The architecture, implementation, and applications of a special-purpose neural network processor are described. The chip performs over 2000 multiplications and additions simultaneously. Its data path is particularly suitable for the convolutional topologies that are typical in classification networks, but can also be configured for fully connected...

A high-speed programmable neural network chip and its application to character recognition are described. A network with over 130000 connections has been implemented on a single chip and operates at a rate of over 1000 classifications per second. The chip performs up to 2000 multiplications and additions simultaneously. Its datapath is suitable for...

Not Available

A board is described that contains the ANNA neural-network chip, anda DSP32C digital signal processor. The ANNA #Analog Neural NetworkArithmetic unit# chip performs mixed analog#digital processing.The combination of ANNA with the DSP allows high-speed, end-toendexecution of numerous signal-processing applications, includingthe preprocessing, the ne...

recent years have seen dramatic development of the scanned-image technologies of facsimile and optical character recognition (OCR). The latest fax machines have higher transmission speeds, higher image resolution, more efficient encoding algorithms, and better error correcting capabilities than predecessors. Such machines provide customers with eff...

The authors outline OCR (optical character recognition) technology developed at AT&T Bell Laboratories, including a recognition network that learns feature extraction kernels and a custom VLSI chip that is designed for neural-net image processing. It is concluded that both high speed and high accuracy can be obtained using neural-net methods for ch...

Hardware architectures for character recognition are discussed, and choices for possible circuits are outlined. An advanced (and working) reconfigurable neural-net chip that mixes analog and digital processing is described. It is found that different approaches to image recognition often lead to neural-net architectures that have limited connectivi...

A large class of applications where theoretical considerations that promote high-accuracy classification result in constrained network architectures have been identified through a series of experiments in pattern recognition using neural net algorithms. These constrained nets can map onto appropriately designed hardware. The concepts learned from t...

It is shown that a neural net can perform handwritten digit recognition with state-of-the-art accuracy. The solution required automatic learning and generalization from thousands of training examples and also required designing into the system considerable knowledge about the task-neither engineering nor learning from examples alone would have suff...

Neural Network research has always interested hardware designers, theoreticians, and application engineers. But until recently, the common ground between these groups was limited: the neural-net chips were too small to implement any full-size application, and the algorithms were too complicated (or the applications not interesting enough) to be imp...

An application of back-propagation networks to handwritten zip code recognition is presented. Minimal preprocessing of the data is required, but the architecture of the network is highly constrained and specifically designed for the task. The input of the network consists of size-normalized images of isolated digits. The performance on zip code dig...

The ability of learning networks to generalize can be greatly enhanced by providing constraints from the task domain. This paper demonstrates how such constraints can be integrated into a backpropagation network through the architecture of the network. This approach has been successfully applied to the recognition of handwritten zip code digits pro...

Two novel methods for achieving handwritten digit recognition are described. The first method is based on a neural network chip that performs line thinning and feature extraction using local template matching. The second method is implemented on a digital signal processor and makes extensive use of constrained automatic learning. Experimental resul...

It is argued that the large interconnectivity and the precision required in neural network models present novel opportunities for analog computing. Analog circuits for a wide variety of problems such as pattern matching, optimization, and learning have been proposed and a few have been built. Most of the circuits built so far are relatively small,...

The authors have demonstrated proximity effect coupling between a high-transition-temperature superconductor and a normal metal. In a device with a 1-μm long gap in a YBa₂Cu₃O₇ film spanned by an Au shunt, the authors observed a DC supercurrent and the AC Josephson effect under microwave irradiation from 2 GHz to 1...

We present an application of back-propagation networks to handwritten digit recognition. Minimal preprocessing of the data was required, but architecture of the network was highly constrained and specifically designed for the task. The input of the network consists of normalized images of isolated digits. The method has 1% error rate and about a 9%...

An abstract is not available.

Early results from exploring alternative computer architectures based on hints from neurobiology suggest that networks of highly-interconnected, simple, low-precision processors can give tools for tackling problems that have been hard or impossible to do on standard computers. The authors describe an electronic neural model and show how this model...

A general-purpose, fully interconnected neural-net chip was used to perform computationally intensive tasks for handwritten digit recognition. The chip has nearly 3000 programmable connections, which can be set for template matching. The templates can be reprogrammed as needed during the recognition sequence. The recognition process proceeds in fou...

We have demonstrated proximity effect coupling between a high transition temperature superconductor and a normal metal. A device with a 1 Am long gold microbridge coupling two evaporated films of YBa2Cu3O7 has a dc supercurrent and exhibits the ac Josephson effect when irradiated with microwaves. These high quality S-N interfaces should have applic...

The authors describe a complementary metal-oxide-semiconductor (CMOS) very-large-scale integrated (VLSI) circuit implementing a connectionist neural-network model. It consists of an array of 54 simple processors fully interconnected with a programmable connection matrix. This experimental design tests the behavior of a large network of processors i...

After discussing possible applications of neural networks, the authors examine hardware implementation including emulators, analog microelectronic implementations, and the synaptic interconnections. Examples of neural network chips are considered. The circuits have programmable interconnections, but the weight values have to be computed externally...

The use of electronic neural networks to handle some complex computing problems is discussed. A simple neural model is shown and discussed in terms of its computational aspects. The use of electronic neural networks in machine pattern recognition and classification and in machine learning is examined. CMOS programmable networks are discussed. 15 re...

We are exploring electronic implementations of fine-grained parallel computing models that are loosely drawn from models of biological neural function. Experimental custom chips that combine a new mix of analog and digital processing with standard fabrication technology have shown the feasibility of the neural network approach. Early results on tra...

. This paper describes a 1MHz CMOS implementation of neural networkused for acoustical attention. Signals are coded into binary spikes modelling the biology.A uniform model of pulse propagating cell is used in different performing stages. Thesignal processing is distributed to neurons and synapses realized as analog circuits.I. IntroductionHandling...

Neural network computing algorithms, which are loosely based on biological models, have been proposed as alternatives to traditional methods for problems in machine perception. The effectiveness of these algorithms largely hinges on whether they can be implemented efficiently on parallel hardware.

This paper reviews two custom electronic circuits that implement some simple models of neural function. The circuits include a thin-film array of read-only resistive synapses and an array of programmable synapses and amplifiers serving as electronic neurons. Circuit performance and architecture are discussed.

Electronic neural networks can perform the function of associative memory. Given an input pattern, the network searches through its stored memories to find which of them best matches the input. Thus the network does a combination of content-addressable search and error correction. The number of random memories that a network can store is limited to...

A high-density matrix of α-Si resistors was made to demonstrate a new type of parallel-processing associative memory consisting of an interconnected array of analog amplifiers. The 22 × 22 resistor matrix was made using a technology compatible with conventional VLSI processing. This demonstration circuit can recall up to four 22- bit memories in 1...

Multiterminal measurements of magnetoresistance fluctuations in silicon inversion-layer nanostructures are extended to probe spacings $L$\ll${}{L}_{$\varphi${}}$, the phase-preserving diffusion length. Unlike for $L>{L}_{$\varphi${}}$, the sizes of the voltage fluctuations are independent of $L$, and have novel correlations consistent with independ...

We have studied the dynamics of a totally interconnected network of nonlinear amplifiers by building model electronic circuits using dense arrays of resistors and discrete amplifiers. Such models have been discussed recently in the context of spin glasses and neural networks. Even without optimization for speed, these circuits easily reproduce and...