Archives for 2016

BioSpaun: A large-scale behaving brain model with complex neurons

Today I read a paper titled “BioSpaun: A large-scale behaving brain model with complex neurons”

The abstract is:

We describe a large-scale functional brain model that includes detailed, conductance-based, compartmental models of individual neurons. We call the model BioSpaun, to indicate the increased biological plausibility of these neurons, and because it is a direct extension of the Spaun model. We demonstrate that including these detailed compartmental models does not adversely affect performance across a variety of tasks, including digit recognition, serial working memory, and counting. We then explore the effects of applying TTX, a sodium channel blocking drug, to the model. We characterize the behavioral changes that result from this molecular level intervention. We believe this is the first demonstration of a large-scale brain model that clearly links low-level molecular interventions and high-level behavior.

Today I read a paper titled “Progressive Gaussian Filtering”

Fuck me! Nice feedback filtering and sampling system!

The abstract is:

In this paper, we propose a progressive Bayesian procedure, where the measurement information is continuously included into the given prior estimate (although we perform observations at discrete time steps). The key idea is to derive a system of ordinary first-order differential equations (ODE) by employing a new coupled density representation comprising a Gaussian density and its Dirac Mixture approximation. The ODE is used for continuously tracking the true non-Gaussian posterior by its best matching Gaussian approximation. The performance of the new filter is evaluated in comparison with state-of-the-art filters by means of a canonical benchmark example, the discrete-time cubic sensor problem. 

Unbounded Human Learning: Optimal Scheduling for Spaced Repetition

Today I read a paper titled “Unbounded Human Learning: Optimal Scheduling for Spaced Repetition”

I like the fact that there is an empirical evaluation of the results against an actual audience and not just merely conjecture. The queuing implementation is useful. My concern is with “human boredom” that could arise from a badly tuned, automatic system (though no worse than a badly tuned system designed by a human). We want to help the subjects memorise what they need, but we also need to keep feeding them novel, i.e. new, knowledge to keep them engaged.

The abstract is:

In the study of human learning, there is broad evidence that our ability to retain information improves with repeated exposure and decays with delay since last exposure. This plays a crucial role in the design of educational software, leading to a trade-off between teaching new material and reviewing what has already been taught. A common way to balance this trade-off is spaced repetition, which uses periodic review of content to improve long-term retention. Though spaced repetition is widely used in practice, e.g., in electronic flashcard software, there is little formal understanding of the design of these systems. Our paper addresses this gap in three ways. First, we mine log data from spaced repetition software to establish the functional dependence of retention on reinforcement and delay. Second, we use this memory model to develop a stochastic model for spaced repetition systems. We propose a queueing network model of the Leitner system for reviewing flashcards, along with a heuristic approximation that admits a tractable optimization problem for review scheduling. Finally, we empirically evaluate our queueing model through a Mechanical Turk experiment, verifying a key qualitative prediction of our model: the existence of a sharp phase transition in learning outcomes upon increasing the rate of new item introductions. 

Design of an Agent for Answering Back in Smart Phones

Today I read a published paper titled “Design of an Agent for Answering Back in Smart Phones”

Shallow fluff with obvious solutions to obvious problems.

The abstract is:

The objective of the paper is to design an agent which provides efficient response to the caller when a call goes unanswered in smartphones. The agent provides responses through text messages, email etc stating the most likely reason as to why the callee is unable to answer a call. Responses are composed taking into consideration the importance of the present call and the situation the callee is in at the moment like driving, sleeping, at work etc. The agent makes decisons in the compostion of response messages based on the patterns it has come across in the learning environment. Initially the user helps the agent to compose response messages. The agent associates this message to the percept it recieves with respect to the environment the callee is in. The user may thereafter either choose to make to response system automatic or choose to recieve suggestions from the agent for responses messages and confirm what is to be sent to the caller.

Study of Gesture Recognition methods and augmented reality

Today I read a paper titled “Study of Gesture Recognition methods and augmented reality”

Difficult to read due to the stilted language but not due to the content. The content of the study was not that revealing in terms of new knowledge. I guess it would be useful if I wasn’t already familiar with the subject and just need a quick overview.

The abstract is:

With the growing technology, we humans always need something that stands out from the other thing. Gestures are most desirable source to Communicate with the Machines. Human Computer Interaction finds its importance when it comes to working with the Human gestures to control the computer applications. Usually we control the applications using mouse, keyboard, laser pointers etc. but, with the recent advent in the technology it has even left behind their usage by introducing more efficient techniques to control applications. There are many Gesture Recognition techniques that have been implemented using image processing in the past. However recognizing the gestures in the noisy background has always been a difficult task to achieve. In the proposed system, we are going to use one such technique called Augmentation in Image processing to control Media Player. We will recognize Gestures using which we are going to control the operations on Media player. Augmentation usually is one step ahead when it comes to virtual reality. It has no restrictions on the background. Moreover it also does not rely on certain things like gloves, color pointers etc. for recognizing the gesture. This system mainly appeals to those users who always looks out for a better option that makes their interaction with computer more simpler or easier.

Unsupervised Incremental Learning and Prediction of Music Signals

Today I read a paper titled “Unsupervised Incremental Learning and Prediction of Music Signals”

Part of some personal research work I am doing with music generation software.

The abstract is:

A system is presented that segments, clusters and predicts musical audio in an unsupervised manner, adjusting the number of (timbre) clusters instantaneously to the audio input. A sequence learning algorithm adapts its structure to a dynamically changing clustering tree. The flow of the system is as follows: 1) segmentation by onset detection, 2) timbre representation of each segment by Mel frequency cepstrum coefficients, 3) discretization by incremental clustering, yielding a tree of different sound classes (e.g. instruments) that can grow or shrink on the fly driven by the instantaneous sound events, resulting in a discrete symbol sequence, 4) extraction of statistical regularities of the symbol sequence, using hierarchical N-grams and the newly introduced conceptual Boltzmann machine, and 5) prediction of the next sound event in the sequence. The system’s robustness is assessed with respect to complexity and noisiness of the signal. Clustering in isolation yields an adjusted Rand index (ARI) of 82.7% / 85.7% for data sets of singing voice and drums. Onset detection jointly with clustering achieve an ARI of 81.3% / 76.3% and the prediction of the entire system yields an ARI of 27.2% / 39.2%.

Combining Brain-Computer Interfaces and Haptics: Detecting Mental Workload to Adapt Haptic Assistance

Today I read a published paper titled “Combining Brain-Computer Interfaces and Haptics: Detecting Mental Workload to Adapt Haptic Assistance”

I think application of an assistance device in the medical field would be mind blowing (no pun intended). A device that could help a tired doctor diagnose a patient or a surgeon complete an operation.

Or in semi-autonomous vehicles when detecting driver fatigue. Hell, even if software development when you are really deep in the code.

The abstract is:

 In this paper we introduce the combined use of Brain-Computer Interfaces (BCI) and Haptic interfaces. We propose to adapt haptic guides based on the mental activity measured by a BCI system. This novel approach is illustrated within a proof-of-concept system: haptic guides are toggled during a path-following task thanks to a mental workload index provided by a BCI. The aim of this system is to provide haptic assistance only when the user’s brain activity reflects a high mental workload. A user study conducted with 8 participants shows that our proof-of-concept is operational and exploitable. Results show that activation of haptic guides occurs in the most difficult part of the path-following task. Moreover it allows to increase task performance by 53% by activating assistance only 59% of the time. Taken together, these results suggest that BCI could be used to determine when the user needs assistance during haptic interaction and to enable haptic guides accordingly. 

Guaranteed bounds on the Kullback-Leibler divergence of univariate mixtures using piecewise log-sum-exp inequalities

Today I read a published paper titled “Guaranteed bounds on the Kullback-Leibler divergence of univariate mixtures using piecewise log-sum-exp inequalities”

Some of the math was a bit obtuse, but a useful method that could be used for fingerprinting malware, tagging and label similar images using information divergence. The fact that this technique does not require a Monte-Carlo simulation is significant.

The abstract is:
The Kullback-Leibler divergence between two mixture models is a core primitive in many signal processing tasks. Since the Kullback-Leibler divergence of mixtures does not admit closed-form formula, it is in practice either estimated using costly Monte-Carlo stochastic integration or approximated using various techniques. We present a fast and generic method that builds algorithmically closed-form lower and upper bounds on the entropy, the cross-entropy and the Kullback-Leibler divergence of mixtures. We illustrate the versatile method by reporting on our experiments for approximating the Kullback-Leibler divergence between univariate exponential mixtures, Gaussian mixtures and Rayleigh mixtures.

Today I read a published paper titled “Virtual Windshields: Merging Reality and Digital Content to Improve the Driving Experience”

We are barely on the cusp of driver assistance devices in vehicles today. If full autonomous vehicles don’t get here first (I suspect they will for 99% of all daily driving), augmented reality and digital windscreens are a game changer.

The abstract is:

In recent years, the use of the automobile as the primary mode of transportation has been increasing and driving has become an important part of daily life. Driving is a multi-sensory experience as drivers rely on their senses to provide them with important information. In a vehicular context human senses are all too often limited and obstructed. Today, road accidents constitute the eighth leading cause of death. The escalation of technology has propelled new ways in which driver’s senses may be augmented. The enclosed aspect of a car, allied with the configuration of the controls and displays directed towards the driver, offer significant advantages for augmented reality (AR) systems when considering the amount of immersion it can provide to the user. In addition, the inherent mobility and virtually unlimited power autonomy transform cars into perfect mobile computing platforms. However, automobiles currently present limited network connectivity and thus the created augmented objects are merely providing information captured by in-vehicle sensors, cameras and other databases. By combining the new paradigm of Vehicular Ad Hoc Networking (VANET) with AR human machine interfaces, we show that it is possible to design novel cooperative Advanced Driver Assistance Systems (ADAS), that base the creation of AR content on the information collected from neighbouring vehicles or roadside infrastructures. As such we implement prototypes of both visual and acoustic AR systems, which can significantly improve the driving experience. We believe our results contribute to the formulation of a vision where the vehicle is perceived as an extension of the body which permeates the human senses to the world outside the vessel, where the car is used as a better, multi-sensory immersive version of a mobile phone that integrates touch, vision and sound enhancements, leveraging unique properties of VANET.

Egocentric vision IT technologies for Alzheimer disease assessment and studies

Today I read a published paper titled “Egocentric vision IT technologies for Alzheimer disease assessment and studies”

Sort of a SenseCam on steroids with applications towards helping Alzheimer’s sufferers. From what I read I believe this work is kind of early days yet, but it could hold great promise.

The abstract is:

Egocentric vision technology consists in capturing the actions of persons from their own visual point of view using wearable camera sensors. We apply this new paradigm to instrumental activities monitoring with the objective of providing new tools for the clinical evaluation of the impact of the disease on persons with dementia. In this paper, we introduce the current state of the development of this technology and focus on two technology modules: automatic location estimation and visual saliency estimation for content interpretation.

Reconstructing Self Organizing Maps as Spider Graphs for better visual interpretation of large unstructured datasets

Today I read a published paper “Reconstructing Self Organizing Maps as Spider Graphs for better visual interpretation of large unstructured datasets”

The abstract is:

Self-Organizing Maps (SOM) are popular unsupervised artificial neural network used to reduce dimensions and visualize data. Visual interpretation from Self-Organizing Maps (SOM) has been limited due to grid approach of data representation, which makes inter-scenario analysis impossible. The paper proposes a new way to structure SOM. This model reconstructs SOM to show strength between variables as the threads of a cobweb and illuminate inter-scenario analysis. While Radar Graphs are very crude representation of spider web, this model uses more lively and realistic cobweb representation to take into account the difference in strength and length of threads. This model allows for visualization of highly unstructured dataset with large number of dimensions, common in Bigdata sources.

Human Pose Estimation from RGB Input Using Synthetic Training Data

Today I read a published paper “Human Pose Estimation from RGB Input Using Synthetic Training Data”

The abstract is:

We address the problem of estimating the pose of humans using RGB image input. More specifically, we are using a random forest classifier to classify pixels into joint-based body part categories, much similar to the famous Kinect pose estimator [11], [12]. However, we are using pure RGB input, i.e. no depth. Since the random forest requires a large number of training examples, we are using computer graphics generated, synthetic training data. In addition, we assume that we have access to a large number of real images with bounding box labels, extracted for example by a pedestrian detector or a tracking system. We propose a new objective function for random forest training that uses the weakly labeled data from the target domain to encourage the learner to select features that generalize from the synthetic source domain to the real target domain. We demonstrate on a publicly available dataset [6] that the proposed objective function yields a classifier that significantly outperforms a baseline classifier trained using the standard entropy objective [10].

Smart Context-aware Rejuvenation of Engagement on Urban Ambient Augmented Things

Today I read a published paper “Smart Context-aware Rejuvenation of Engagement on Urban Ambient Augmented Things”

The abstract is:

The concern over global urbanization trend imposes smart-city as enabling information and communication technology (ICT) to improve urban governance. However, the light trance on better living space is stimulated by socioeconomic impact of escalated senior generation. Hence, ambient assisted living (AAL) emerges for the autonomous provisioning of pervasive things or objects from relevant perturbation for advanced scientific instrumentation. Meanwhile, citizens are observed in being transfixed by lively stimuli of monotonous urban events with the advent of virtual reality or augmented things. Thus, due to the involvement of situation-awareness or contextualization, engagement/participation information as a utility promises to improve urban experience. However, it is complex to grapple meaningful concepts due to personalization obstacles, such as citizen psychology, information gap, service-visualization. Moreover, recommended practices deficit adaptation to monochromatic choice, disparate impairments, mobility and annotation-richness in urban space. Hence, rejuvenation of engagement relates to monitoring and quantification of ‘service consumption and graceful degradation’ of experience. However, paramount challenges are imposed on this stipulation, such as, unobservability, independence and composite relationship of contexts. Therefore, a parametric Bayesian based model is envisioned to address observability and scalability of contexts and its conjugal relationship with engagement. Last but not the least, systematic framework is demonstrated, which pinpoints key goals of context-aware engagement from participants’ opinions, usages and feed-backs.