김지훈 교수

This paper describes the energy-efficient implementation of a high performance parallel Radix-4 turbo decoder, which is designed to support multiple fourth-generation (4G) wireless communication standards such as mobile WiMAX and 3GPP-LTE. We propose a new hardware architecture that can share hardware resources for the two standards. It mainly consists of eight retimed Radix-4 soft-input soft-output (SISO) decoders to achieve high throughput and a dual-mode parallel hardware interleaver to suppo

We present an energy-efficient processing-in-memory (PIM) architecture named Z-PIM that supports both sparsity handling and fully variable bit-precision in weight data for energy-efficient deep neural networks. Z-PIM adopts the bit-serial arithmetic that performs a multiplication bit-by-bit through multiple cycles to reduce the complexity of the operation in a single cycle and to provide flexibility in bit-precision. To this end, it employs a zero-skipping convolution SRAM, which performs in-mem

The all-solution-based memory devices demonstrated excellent bipolar switching behavior with a high resistive switching ratio of 10³, excellent endurance of more than 1000 cycles, stable retention time greater than 10⁴s at elevated temperatures, and fast programming speed of 250 ns.

Ras and Rho small GTPases are critical for numerous cellular processes including cell division, migration, and intercellular communication. Despite extensive efforts to visualize the spatiotemporal activity of these proteins, achieving the sensitivity and dynamic range necessary for in vivo application has been challenging. Here, we present highly sensitive intensiometric small GTPase biosensors visualizing the activity of multiple small GTPases in single cells in vivo. Red-shifted sensors combi

This brief presents an energy-efficient soft-input soft-output (SISO) decoder based on border metric encoding, which is especially suitable for nonbinary circular turbo codes. In the proposed method, the size of the branch memory is reduced to half and the dummy calculation is removed at the cost of a small-sized memory that holds encoded border metrics. Due to the infrequent accesses to the border memory and its small size, the energy consumed for SISO decoding is reduced by 26.2%. Based on the

Nonbinary turbo codes have many advantages over single-binary turbo codes, but their decoder implementations require much more memory, particularly for storing symbolic extrinsic information to be exchanged between two soft-input-soft-output (SISO) decoders. To reduce the memory size required for double-binary turbo decoding, this paper presents a new method to convert symbolic extrinsic information to bit-level information and vice versa. By exchanging bit-level extrinsic information, the numbe

<inline-formula><tex-math notation="LaTeX">$K$</tex-math></inline-formula> -nearest neighbor search is one of the fundamental tasks in various applications and the hierarchical navigable small world (HNSW) has recently drawn attention in large-scale cloud services, as it easily scales up the database while offering fast search. On the other hand, a computational storage device (CSD) that combines programmable logic and storage modules on a single board becomes popular to address the data bandwid

An ultra-low-power analog-digital hybrid always-on face recognition (FR) processor integrated with a CMOS image sensor (CIS) is proposed for the wearable mobile devices applications such as user authentication. The proposed processor is the first IC with full process of FR in a single chip. The processor adopts analog-digital hybrid convolution operation for efficient integration of CNN processor with CIS. The analog convolution processor is proposed for the computation of the 1 <sup xmlns:mml="

This paper presents Z-PIM, an energy-efficient processing-in-memory (PIM) architecture that supports zero-skipping operations and fully-variable weight bit-precision for efficient deep neural network (DNN). The 8T-SRAM cell based bit-serial operation with hierarchical bit-line structure enables variable weight precision and reduces bit-line switching by 95.42% in convolution layers of VGG-16. Z-PIM handles abundant zeros in weight data by skip-reading their corresponding input data while read-se

Motion in-betweening (MIB) is a process of generating intermediate skeletal movement between the given start and target poses while preserving the naturalness of the motion, such as periodic footstep motion while walking. Although state-of-the-art MIB methods are capable of producing plausible motions given sparse key-poses, they often lack the controllability to generate motions satisfying the semantic contexts required in practical applications. We focus on the method that can handle pose or s

Differential phase optical coherence tomography (DPOCT) is introduced to measure optical path length changes in response to pulsed laser irradiation (585 nm). An analytical equation that includes thermoelastic surface displacement and thermorefractive index change is derived to predict optical path length change in response to pulsed laser irradiation for both "confined surface" and "free surface" model systems. The derived equation is tested by comparing predicted values with data recorded from

A 50 Mbps, 2.24 mm <sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</sup> double-binary turbo double decoder is designed and implemented in 0.13 mum CMOS process for the WiMAX standard. To reduce the large extrinsic memory needed in double-binary turbo decoding, the proposed decoder exchanges the bit-level extrinsic information values rather than the traditional symbol-level extrinsic information values, which is achieved by deriving two simple con

Discovery of the naturally evolved fluorescent proteins and their genetically engineered biosensors have enormously contributed to current bioimaging techniques. These reporters to trace dynamic changes of intracellular protein activities have continuously transformed according to the various demands in biological studies. Along with that, light-inducible optogenetic technologies have offered scientists to perturb, control and analyze the function of intracellular machineries in spatiotemporal m

An energy harvesting system using a triboelectric generator (TEG), which converts a small amount of mechanical energy to available electrical energy, has recently been developed by combining a simple one-directional mechanical force (contact and separation or sliding back and forth) with a 2D device materials. However, with regard to using the TEG in real world applications, there is no TEG design suitable for utilizing a variety of mechanical forces and for generating triboelectric charge in va

Although recent works on neural vocoder have improved the quality of synthesized audio, there still exists a gap between generated and ground-truth audio in frequency space. This difference leads to spectral artifacts such as hissing noise or reverberation, and thus degrades the sample quality. In this paper, we propose Fre-GAN which achieves frequency-consistent audio synthesis with highly improved generation quality. Specifically, we first present resolution-connected generator and resolution-