MOTIVATION: Lipids play crucial roles in various biological functions and diseases. However, a gap exists in databases providing information of lipids functions based on curated information. Consequently, LipidFun is purposed as the first lipid function database with sentence-level evidence detailing lipid-related phenotypes and biological functions. RESULTS: Potential lipid functions were extracted from the biomedical literature using natural language processing techniques, with accuracy and reliability ensured through manual curation by four domain experts. LipidFun constructs classification systems for lipids, biological functions, and phenotypes for named entity recognition. Sentence-level evidence is extracted to highlight connections to lipid-associated biological processes and diseases. Integrating these classification systems and a large amount of sentence-level evidence allows LipidFun to provide an overview of lipid-phenotype and lipid-biological function associations through concise visualizations. Overall, LipidFun unravels the relationships between lipids and biological mechanisms, underscoring their overarching influence on physiological processes. AVAILABILITY AND IMPLEMENTATION: LipidFun is available at https://lipidfun.bioinfomics.org/.

A latent style model describing manga styles based on the proposed manga-specific features is constructed to facilitate novel style-based applications. Two manga-specific features, i.e., screentone features showing texture and shade, and panel features showing panel arrangement, are firstly proposed to describe manga pages. Based on the latent Dirichlet allocation technique, we discover latent style elements embedded in manga documents, which are described by visual words derived from manga-specific features. Distributions of style elements are then used to measure similarity between manga documents, and facilitate the development ofvarious style-based applications. Experimental results show that the features and models especially designed for describing manga styles yield promising performance and could bring many potential extensions.

A wide-locking range divide-by-3 LC injection-locked frequency divider (ILFD) is proposed and implemented in the TSMC 0.18-μm 1P6M CMOS process. The divide-by-3 ILFD uses a cross-coupled nMOS pair, an injection MOSFET pair, and a sixth-order RLC resonator. The divide-by-3 ILFD has three oscillation frequency bands and three locking ranges. At the drain-source bias VDDof 0.9 V and at the incident power of 0 dBm, by switching a varactor bias, the high-band, middle-band, and low-band locking ranges of the ILFD are, respectively, given by 3.47 GHz (33.87%) from 8.51 to 11.98 GHz, 3.72 GHz (54.86%) from 4.92 to 8.64 GHz, and 1.05 GHz (19.76%) from 4.88 to 5.93 GHz. The operation range is 7.1 GHz (84.2%) from 4.88 to 11.98 GHz. The ILFD-core power consumption is 6.759 mW. A wider locking range of 4.43 GHz (67.07%) from 4.39 to 8.82 GHz is obtained by overlapping the three locking ranges at a fixed bias condition. The die area is 0.991 × 1.04 mm2.

BACKGROUND: In the application of microarray data, how to select a small number of informative genes from thousands of genes that may contribute to the occurrence of cancers is an important issue. Many researchers use various computational intelligence methods to analyzed gene expression data. RESULTS: To achieve efficient gene selection from thousands of candidate genes that can contribute in identifying cancers, this study aims at developing a novel method utilizing particle swarm optimization combined with a decision tree as the classifier. This study also compares the performance of our proposed method with other well-known benchmark classification methods (support vector machine, self-organizing map, back propagation neural network, C4.5 decision tree, Naive Bayes, CART decision tree, and artificial immune recognition system) and conducts experiments on 11 gene expression cancer datasets. CONCLUSION: Based on statistical analysis, our proposed method outperforms other popular classifiers for all test datasets, and is compatible to SVM for certain specific datasets. Further, the housekeeping genes with various expression patterns and tissue-specific genes are identified. These genes provide a high discrimination power on cancer classification.

Contemporary cache is known for consuming a large part of total power in microprocessors. Voltage scaling had been used to reduce the power consumption of the cache. However, due to the impact of variations, SRAM cells of the cache could potentially fail when voltage dropping. To against variations, we need to increase the supply voltage for the safety margin, thus the cache costs large energy consumption. For eliminating the voltage safety margin, some prior works for SRAM failure tolerance designs were proposed. These schemes will result in worse energy consumption and cannot deal with dynamic variations. They still have a safety margin to resist dynamic variations. With the supply voltage scaling down, we find out that the major reason of failures is that some slow cells have longer latency. We call these cell faults as “latency fault”. If each cache line can be accessed in an appropriate access time, the slower cells could be reused but not disable them. We propose a VAL-Cache adapting the access time to tolerate latency faults and which is able to scale down the voltage. And we also propose the latency-fault detector to detect latency faults at run-time so as to tolerate both static and dynamic variations. Our experimental results on Mibench and 0xbench benchmarks demonstrate that the energy consumption can be reduced 10%~18% in average at a cost of acceptable performance loss.

BACKGROUND: Over the past decade, gene expression microarray studies have greatly expanded our knowledge of genetic mechanisms of human diseases. Meta-analysis of substantial amounts of accumulated data, by integrating valuable information from multiple studies, is becoming more important in microarray research. However, collecting data of special interest from public microarray repositories often present major practical problems. Moreover, including low-quality data may significantly reduce meta-analysis efficiency. RESULTS: M2DB is a human curated microarray database designed for easy querying, based on clinical information and for interactive retrieval of either raw or uniformly pre-processed data, along with a set of quality-control metrics. The database contains more than 10,000 previously published Affymetrix GeneChip arrays, performed using human clinical specimens. M2DB allows online querying according to a flexible combination of five clinical annotations describing disease state and sampling location. These annotations were manually curated by controlled vocabularies, based on information obtained from GEO, ArrayExpress, and published papers. For array-based assessment control, the online query provides sets of QC metrics, generated using three available QC algorithms. Arrays with poor data quality can easily be excluded from the query interface. The query provides values from two algorithms for gene-based filtering, and raw data and three kinds of pre-processed data for downloading. CONCLUSION: M2DB utilizes a user-friendly interface for QC parameters, sample clinical annotations, and data formats to help users obtain clinical metadata. This database provides a lower entry threshold and an integrated process of meta-analysis. We hope that this research will promote further evolution of microarray meta-analysis.

High power consumption and limited viewing angles are the two major drawbacks of liquid crystal displays (LCDs). This paper models the angular-dependent luminance attenuation of a twisted nematic LCD and incorporates user's viewing direction into a backlight scaling framework such that the power consumption can be minimized while preserving the optimal image quality. We also studied the human factors of viewing direction variation, and developed an emulator for visualizing luminance attenuation and color shift in different viewing directions.

This paper presents an algorithm for minimizing power consumption of LED backlights in transmissive TFT-LCD monitors. The proposed algorithm reduces power consumption by scaling the luminous intensity of the red, green, and blue LED backlights independently according to the image histograms of each color channel. The algorithm consists of two phases. The first phase, chromaticity scaling, finds the optimal ratios of red, green, and blue backlights subject to a perceived color difference constraint. The second phase, luminance scaling, finds the optimal dimming factor subject to a perceived lightness difference constraint. The perceived color and lightness differences are measured by the CIELAB Color Difference Equation 2Δ E, a standard metric for measuring color ab variation. Psychophysical experiments were performed with 35 observers to uncover the optimal luminance scaling function. An experimental LED backlight module was implemented and installed on a 19 ” side-lit TFT-LCD monitor to replace the original CCFL backlight. Within limited perceivable difference 2Δ E, up to 76% ab of power consumption can be reduced for the benchmark images.

This paper presents an algorithm for minimizing power consumption of backlights in color sequential displays. The proposed algorithm reduces power consumption by scaling the luminous intensity of the red, green, and blue backlights independently according to the image histograms of each color field. The algorithm consists of two phases. The first phase, chromaticity scaling, finds the optimal ratios of red, green, and blue backlights subject to a perceived color difference constraint. The second phase, luminance scaling, finds the optimal dimming factor subject to a perceived lightness difference constraint. The perceived color and lightness differences are measured by the CIELAB Color Difference Equation ΔE*ab, a standard metric for measuring color variation. The proposed method can be applied to RGB LED-backlit LCDs as well. Within limited perceivable difference ΔE*ab ≤ 2, up to 64% of power consumption can be reduced for the benchmark images.

This paper presents a novel backlight driving technique for liquid crystal displays. By scaling the intensity, frequency, and duty cycle of the backlight, this technique not only increases the perceived brightness but also prolongs the service time of rechargeable batteries. The increased brightness comes from a perceptual effect of temporal vision - a brief flash appears brighter than a steady light of the same intensity, called Brücke brightness enhancement effect. The prolonged service time comes from the relaxation phenomenon - a lithium-ion battery lasts longer by pulsed discharge. Combining these two effects, a great amount of service time can be obtained at the cost of flickering. We performed visual experiments to parameterize the Brücke effect and derived an optimization algorithm accordingly. To demonstrate the potential energy savings of this technique, we profiled the power consumption of an Apple iPod and fabricated an LED driving module. Based on experimental data, 75% of energy consumption can be saved and the service time can be extended to 300%.

This paper presents a concurrent brightness and contrast scaling (CBCS) technique for a cold cathode fluorescent lamp (CCFL) backlit TFT-LCD display. The proposed technique aims at conserving power by reducing the backlight illumination while retaining the image fidelity through preservation of the image contrast. First, we explain how CCFL works and show how to model the non-linearity between its backlight illumination and power consumption. Next, we propose the contrast distortion metric to quantify the image quality loss after backlight scaling. Finally, we formulate and optimally solve the CBCS optimization problem with the objective of minimizing the fidelity and power metrics. Experimental results show that an average of 3.7X power saving can be achieved with only 10% of contrast distortion.

This paper describes a dynamic voltage and frequency scaling (DVFS) technique for MPEG decoding to reduce the energy consumption while maintaining a quality of servic(QoS) constraint. The computational workload for an incoming frame is predicted using a frame-based history so that the processor voltage and frequency can be scaled to provide the exact amount of computing power needed to decode the frame. More precisely, the required decoding time for each frame is separated into two parts: a frame-dependent (FD) part and a frame-independent (FI) part. The FD part varies greatly according to the type of the incoming frame whereas the FI part remains constant regardless of the frame type. In the DVFS scheme presented in this paper the FI part is used to compensate for the prediction error that may occur during the FD part such that a significant amount of energy can be saved while meeting the frame rate requirement. The proposed DVFS algorithm has been implemented on a StrongArm-1110 based evaluation board. Measurement results demonstrate a higher than 50% CPU energy saving as a result of DVFS.

This paper presents an irredundant encoding technique to minimize the switching activity on a multiplexed dynamic RAM (DRAM) address bus. The DRAM switching activity can be classified either as external (between two consecutive addresses) or internal (between the row and column addresses of the same address). For external switching activity in a sequential access pattern, we present a power-optimal encoding, named Pyramid code. Extensions of the basic code address different types of DRAM devices. The proposed codes reduce power dissipation on the memory bus by a factor of two or more.

This paper presents encoding techniques to optimize the switching activity on a multiplexed DRAM address bus. The DRAM switching activity can be classified either as external (between two consecutive addresses) or internal (between the row and column addresses of the same address). To eliminate the external switching activity for sequential access, we propose an optimal encoding, Pyramid code, for conventional DRAM mode as well as Burst Pyramid code for burst mode DRAM. To minimize the internal switching activity, we propose Scattered Paging for both random and sequential access patterns by exploiting the built-in virtual memory mechanism, which is commonly present on modern processors.

This paper presents Pyramid code, an optimal code for transmitting sequential addresses over a DRAM bus. Constructed by finding an Eulerian cycle on a complete graph, this code is optimal for conventional DRAM in the sense that it minimizes the switching activity on the time-multiplexed address bus from CPU to DRAM. Experimental results on a large number of testbenches with different characteristics (i.e. sequential vs. random memory access behaviors) are reported and demonstrate a reduction of bus activity by as much as 50%.