자연주의적 탐구조사(Naturalistic Inquiry)

Machine Learning is the study of methods for programming computers to learn. Computers are applied to a wide range of tasks, and for most of these it is relatively easy for programmers to design and implement the necessary software. However, there are many tasks for which this is difficult or impossible. These can be divided into four general categories. First, there are problems for which there exist no human experts. For example, in modern automated manufacturing facilities, there is a need to predict machine failures before they occur by analyzing sensor readings. Because the machines are new, there are no human experts who can be interviewed by a programmer to provide the knowledge necessary to build a computer system. A machine learning system can study recorded data and subsequent machine failures and learn prediction rules. Second, there are problems where human experts exist, but where they are unable to explain their expertise. This is the case in many perceptual tasks, such as speech recognition, hand-writing recognition, and natural language understanding. Virtually all humans exhibit expert-level abilities on these tasks, but none of them can describe the detailed steps that they follow as they perform them. Fortunately, humans can provide machines with examples of the inputs and correct outputs for these tasks, so machine learning algorithms can learn to map the inputs to the outputs. Third, there are problems where phenomena are changing rapidly. In finance, for example, people would like to predict the future behavior of the stock market, of consumer purchases, or of exchange rates. These behaviors change frequently, so that even if a programmer could construct a good predictive computer program, it would need to be rewritten frequently. A learning program can relieve the programmer of this burden by constantly modifying and tuning a set of learned prediction rules. Fourth, there are applications that need to be customized for each computer user separately. Consider, for example, a program to filter unwanted electronic mail messages. Different users will need different filters. It is unreasonable to expect each user to program his or her own rules, and it is infeasible to provide every user with a software engineer to keep the rules up-to-date. A machine learning system can learn which mail messages the user rejects and maintain the filtering rules automatically. Machine learning addresses many of the same research questions as the fields of statistics, data mining, and psychology, but with differences of emphasis. Statistics focuses on understanding the phenomena that have generated the data, often with the goal of testing different hypotheses about those phenomena. Data mining seeks to find patterns in the data that are understandable by people. Psychological studies of human learning aspire to understand the mechanisms underlying the various learning behaviors exhibited by people (concept learning, skill acquisition, strategy change, etc.).

Machine learning

https://link.springer.com/content/pdf/10.2307%2F3343168.pdf

Development as Freedom

The goal of this article is to review the state-of-the-art tracking methods, classify them into different categories, and identify new trends. Object tracking, in general, is a challenging problem. Difficulties in tracking objects can arise due to abrupt object motion, changing appearance patterns of both the object and the scene, nonrigid object structures, object-to-object and object-to-scene occlusions, and camera motion. Tracking is usually performed in the context of higher-level applications that require the location and/or shape of the object in every frame. Typically, assumptions are made to constrain the tracking problem in the context of a particular application. In this survey, we categorize the tracking methods on the basis of the object and motion representations used, provide detailed descriptions of representative methods in each category, and examine their pros and cons. Moreover, we discuss the important issues related to tracking including the use of appropriate image features, selection of motion models, and detection of objects.

/pdf/object-tracking-a-survey-3706dclxb5.pdf

Object tracking: A survey

In this paper, we propose a novel explicit image filter called guided filter. Derived from a local linear model, the guided filter computes the filtering output by considering the content of a guidance image, which can be the input image itself or another different image. The guided filter can be used as an edge-preserving smoothing operator like the popular bilateral filter [1], but it has better behaviors near edges. The guided filter is also a more generic concept beyond smoothing: It can transfer the structures of the guidance image to the filtering output, enabling new filtering applications like dehazing and guided feathering. Moreover, the guided filter naturally has a fast and nonapproximate linear time algorithm, regardless of the kernel size and the intensity range. Currently, it is one of the fastest edge-preserving filters. Experiments show that the guided filter is both effective and efficient in a great variety of computer vision and computer graphics applications, including edge-aware smoothing, detail enhancement, HDR compression, image matting/feathering, dehazing, joint upsampling, etc.

/pdf/guided-image-filtering-5atmpsksoa.pdf

Guided Image Filtering

We present Collage, a software application designed for classroom presentations for school teachers in the developing world. An in-depth investigation of teaching practices in several schools in India led us to believe that a simple tool that enabled the display of images and textbook materials while facilitating blackboard-like interactions would be very helpful for these teachers. Collage is a simple media viewer with a small number of features that enables teachers to prepare lessons with little overhead and then present them in classrooms with maximum flexibility. The tool was piloted in three schools in suburban India for use in real-world classroom teaching. All teachers who used Collage uniformly praised it, and students' learning of visual concepts seemed to improve through it. Interestingly, some teachers continue to use the tool now for their own teaching needs and have spontaneously shared it with colleagues from other schools.

/pdf/collage-a-presentation-tool-for-school-teachers-1fgwyapsvq.pdf

Collage: a presentation tool for school teachers

HCI for Development (HCI4D) lies at the intersection of information communication technologies for development (ICT4D) and human-computer interaction (HCI). The mainstream HCI community creates user experiences for the developed-world consumer, while ICT4D is concerned about creating relevant technologies for developing nations. The fusion—HCI4D—evolved and realigned goals to design user experiences for a new audience, namely populations living in a context of low rates of telecom diffusion and digital literacy. The foundation of good interaction design is understanding the user [1]. While usable interfaces are critical for good user experience, contex-tual factors such as institutional arrangements, literacy levels, and social, political, economic, and infrastructural issues often guide the usage and sustainability of development projects. In this regard, ethnography is a highly favored field technique in HCI4D research. This is due to the perspective it lends in gauging the sociocultural relevance and acceptability of technologies in a given context. At the core of ethnographic research is field immersion of the researcher as a participant-observer. It follows that ethnographic studies are not only vulnerable to biases held by the researcher, but also are products of relationships established between the researcher and infor-mants. Ethnography has historically involved power imbalances between researcher and infor-mants. In the context of HCI4D, projects may fall into the trap of mistranslating findings into a design irrelevant to the needs of target users in specific socioeconomic contexts, even with the best of intentions. Misreading cultures can disrupt the developmental underpinnings of HCI4D, which is concerned with technologies that move toward fulfilling human developmental goals. From our own experience in employing ethnographic methods in HCI4D, to avoid these traps: Rearticulate the assumptions of developmen-1. tal projects, Outline a repertoire of field techniques 2. improvised for low-income settings, and Highlight lessons learned from tensions 3. rooted in the conflicting cultural contexts of the HCI4D researcher and informants. Our ideas developed out of field engagements between 2008 and 2009 in the slums of Bangalore, India, where we observed female domestic workers , and Mumbai, where we studied small businesses and their socioeconomic networks [2]. While our reflections are not new to the field of anthropology , we present ways to manage these age-old problems in the context of HCI4D. Some salient observations from our fieldwork are as follows: Question the notion of development. Understanding the meaning of development is critical to any developmental project. Broadly speaking, development as a goal addresses the necessities …

UNDER DEVELOPMENTEncountering development ethnographically

In anticipation of the various and imminent limits to global consumption-based growth, some quarters of the technology industry have begun to consider the novel use of computing to prevent, postpone, alleviate, or recover from a crisis in what could be called “sustainable computing.” There is, however, a great danger that certain inclinations of the technology sector will undermine the very goals of sustainability and resilience that such efforts claim to seek. This paper outlines a preliminary taxonomy of value for sustainable computing projects. The taxonomy suggests a three-dimensional classification of projects in terms of their impact on, intention toward, and effort required for sustainability. By making explicit an evaluative framework by which computing might and might not contribute to sustainability goals, the hope is that future work will tend toward projects that are genuinely helpful.

Preliminary thoughts on a taxonomy of value for sustainable computing

We present research leading toward an understanding of the optimal audio-visual representation for illustrating concepts for illiterate and semi-literate users of computers. In our user study, which to our knowledge is the first of its kind, we presented each of 13 different health symptoms to 200 illiterate subjects in one representation randomly selected among the following ten: text, static drawings, static photographs, hand-drawn animations, and video, each with and without voice annotation. The goal was to see how comprehensible these representation types were for an illiterate audience. We used a methodology for generating each of the representations tested in a way that fairly stacks one representational type against the others. Our main results are that (1) richer information is not necessarily better understood overall; (2) voice annotation generally helps in speed of comprehension, but bimodal audio-visual information can be confusing for the target population; (3) the relative value of dynamic imagery versus static imagery depends on other factors. Analysis of these statistically significant results and additional detailed results are also provided.

/pdf/optimal-audio-visual-representations-for-illiterate-users-3hrg03rnvw.pdf

Optimal audio-visual representations for illiterate users

In the United States, African-American women are three times more likely to die of pregnancy-related issues, and 57% more likely to have miscarriages than White women In part, this is due to deficiencies in the information ecology serving pregnant mothers We report on a qualitative study with 16 low-income, African-American, pregnant women from the Detroit area, focused on understanding their informational needs and information-seeking behavior We find that our participants consumed information from multiple sources voraciously; treated formal medical sources as no more trustworthy than other information sources; sought corroborative evidence when making decisions; relied on video and social media sources; and voiced requests for material and social support more than for medical information Among our recommendations are increased use of video as a means for pregnancy-related information dissemination -- particularly testimonials by experienced mothers with similar backgrounds -- and better integration of medically sound voices in spaces where mothers already congregate online

Kentaro Toyama

Papers

Collage: a presentation tool for school teachers

UNDER DEVELOPMENTEncountering development ethnographically

Preliminary thoughts on a taxonomy of value for sustainable computing

Optimal audio-visual representations for illiterate users

An Exploration of African-American Pregnant Women's Information-Seeking Behavior in Detroit