This Research-to-Practice Work In Progress paper presents a combination of curricular analytics and student performance analysis in a Bachelor of Science in Electrical Engineering (BSEE) degree program. The study of curricular complexity has gained attention in recent years as a means to quantify the ease of progression through a degree program and to highlight bottleneck courses. While there is much attention to canonical curricular complexity of a recommended degree roadmap, there is little analysis of what paths students actually take through a curriculum. Here, we apply the concept of curricular complexity to students' progression through a BSEE degree and compare those results to the curricular complexity of the canonical degree roadmap. We study the relationship between time-to-graduation, student schedule complexity, and number of extraneous courses. We find that many students take breaks during their degree, extending time-to-graduation, even while many enter the degree with transfer credits. We find a direct relationship between schedule complexity and time-to-graduation, with a one-term increase in time-to-graduation similar to but slightly lower than the average term-weighted complexity. Finally, we find that students take extraneous courses consistent with three to four extra terms, indicating an additional factor in time-to-graduation.

This Research-to-Practice Work in Progress paper presents an analysis of student performance in courses in light of the prerequisite structure of those courses and the degree of topic-level dependencies between the courses. Prerequisite relationships, especially long sequences of relationships, are the largest contributing factor to curricular complexity, a measure of the ease of progression through a degree. While the effect of prerequisites on curricular complexity can be quantitatively measured, the relationship of student performance between the courses in a course sequence and the topic-level dependency between the courses have not been studied. Here we study the correlation in course grades in an electrical engineering curriculum. Additionally, we quantify the topic-level dependencies between courses by building a hierarchy of topics through the electrical engineering curriculum. These preliminary analyses indicate that there are likely other strong factors that affect class grades beyond specific course competencies and that the prerequisite structure of the curriculum may not necessarily accurately reflect a relationship in terms of course competencies.

This Innovative Practice Category Work In Progress paper presents an application of machine learning and data mining to student performance data in an undergraduate electrical engineering program. We are developing an analytical approach to enhance retention in the program especially among underrepresented groups. Our approach will provide quantitative assessment of student performance in courses. Specifically, by hierarchically mapping the content of assignments to course learning objectives, we can better decipher which concepts a particular student is struggling with and, with the help of peer mentors, create tailored intervention techniques to help the student be successful in the program. These results will also be useful to academic advisors who can work with the student to determine class schedules that promote success in the program. In addition, students can take a proactive approach to their learning. In our approach, data from our learning management system and other available sources will be used to predict several outcomes for individuals such as when a student is beginning to have trouble with the material or if factors outside of the classroom are affecting their success. Here, we present our initial database schema and preliminary results relating number of class re-takes to time-to-graduation.

This Innovative Practice Category full paper presents research in which Entrepreneurial Minded Learning (EML) was implemented in introductory engineering classes, as an intervention, to study how EML influences self-efficacy between students of Hispanic and non-Hispanic ethnicities at a Hispanic Serving Institution. The EML tenet is to include the “three C's” in project-based learning: curiosity, connections, and creating value. A key question is whether EML not only influences student learning but also the students' confidence in being successful engineers. Much work has been done studying how EML affects learning, yet few studies look into the effect of the mindset on self-efficacy. Moreover, self-efficacy is known to differ depending on a person's race and ethnicity. We studied three groups: (i) a cohort not exposed to EML; (ii) a cohort exposed to EML design projects lightly guided by a student-mentor; (iii) a cohort exposed to EML design projects heavily guided by faculty and a graduate student. We performed pre and post-surveys to evaluate how strongly students identify as engineers when exposed (or not) to EML. We find that participation in an introductory engineering course itself may have more influence on students' engineering self-efficacy than the specific level of EML within the course.

In this paper, we propose a two-stage speech enhancement technique. In the training stage, a Gaussian Mixture Model (GMM) of the mel-frequency cepstral coefficients (MFCCs) of a user’s clean speech is computed wherein the component densities of the GMM serve to model the user’s “acoustic classes.” In the enhancement stage, MFCCs from a noisy speech signal are computed and the underlying clean acoustic class is identified via a maximum a posteriori (MAP) decision and a novel mapping matrix. The associated GMM parameters are then used to estimate the MFCCs of the clean speech from the MFCCs of the noisy speech. Finally, the estimated MFCCs are transformed back to a time-domain waveform. Our results show that we can improve PESQ in environments as low as−10 dB SNR.

In this paper, we propose a low bit-rate speech codec based on vector quantization (VQ) of the mel-frequency cepstral coefficients (MFCCs). We begin by showing that if a high-resolution mel-frequency cepstrum (MFC) is computed, good-quality speech reconstruction is possible from the MFCCs despite the lack of phase information. By evaluating the contribution toward speech quality that individual MFCCs make and applying appropriate quantization, our results show that the MFCC-based codec exceeds the state-of-the-art MELPe codec across the entire range of 600-2400 bps, when evaluated with the perceptual evaluation of speech quality (PESQ) (ITU-T recommendation P.862). The main advantage of the proposed codec is in distributed speech recognition (DSR) since the MFCCs can be directly applied thus eliminating additional decode and feature extract stages; furthermore, the proposed codec better preserves the fidelity of MFCCs and better word accuracy rates as compared to CELP and MELPe codecs.

In this paper, we propose a low bit-rate speech codec based on a hybrid scalar/vector quantization of the mel-frequency cepstral coefficients (MFCCs). We begin by showing that if a high-resolution mel-frequency cepstrum (MFC) is computed, good-quality speech reconstruction is possible from the MFCCs despite the lack of explicit phase information. By evaluating the contribution toward speech quality that individual MFCCs make and applying appropriate quantization, our results show perceptual evaluation of speech quality (PESQ) of the MFCC-based codec matches the state-of-the-art MELPe codec at 600 bps and exceeds the CELP codec at 2000-4000 bps coding rates. The main advantage of the proposed codec is in distributed speech recognition (DSR) since speech features based on MFCCs can be directly obtained from code words thus eliminating additional decode and feature extract stages.

Many features used in the analysis of pathology imagery are inspired by grading features defined by clinical pathologists as important for diagnosis and characterization. A large majority of these features are features of cell nuclei; as such, there is often the desire to segment the imagery into individual nuclei prior to feature extraction and further analysis. In this paper we present an analysis of the utility of imperfectly segmented cell nuclei for classification of H&E stained histopathology imagery of breast tissue. We show the object- and image-level classification performance using these imperfectly segmented nuclei in a benign versus malignant decision. Results indicate that very good classification accuracies can be achieved with imperfectly segmented nuclei and further that perfect nuclei segmentation does not necessarily guarantee better classification accuracy.

We consider here the implementation of efficient search and retrieval of digital elevation maps (DEMs), in particular the ability to conduct elevation searches without decompressing the entire map. We utilize set partitioning in hierarchical trees (SPIHT) for the compression, and compare two wavelet-based search and retrieval systems: one utilizing nonlinear max- and min-lifted integer wavelets and the other using the standard 5/3 integer wavelet. While the coarse-scale maxima/minima preservation inherent in the max- and min-lifted wavelets may seem ideal for the implementation of efficient elevation searches, the interscale propagation of maxima/minima and the bit-plane layering of the SPIHT coder adversely affect the efficiency of such searches. The use of the standard 5/3 integer wavelet requires the addition of separate maxima and minima information but does not have the complication of coefficient propagation, providing more satisfactory results. Elevation search bitrates are presented for both systems and compared to results obtained using the Kakadu implementation of the JPEG2000 standard.

In this paper we consider the use of nonlinear max- and min-lifted integer wavelets encoded using set partitioning in hierarchical trees (SPIHT) for the compression and efficient data-retrieval of digital elevation maps (DEMs). Max- and min-lifted integer wavelets losslessly code data such that the regional maxima and minima are preserved at the coarsest scale, which facilitates efficient multiresolutional elevation searches. Of particular interest in this application are the effects of maxima/minima propagation between resolution scales, as well as the effects of the bit-plane layering of wavelet coefficients inherent in the SPIHT coder. This implementation of a DEM browsing system will be compared to one utilizing a standard 5/3 wavelet, such as is used in JPEG2000.