Jong‐Ho Lee

Adrian de Wynter, Ishaan Watts, Tua Wongsangaroonsri et al. · cmu, deepmind

Large language models (LLMs) and small language models (SLMs) are being adopted at remarkable speed, although their safety still remains a serious concern. With the advent of multilingual S/LLMs, the question now becomes a matter of scale: can we expand multilingual safety evaluations of these models with the same velocity at which they are deployed? To this end, we introduce RTP-LX, a human-transcreated and human-annotated corpus of toxic prompts and outputs in 28 languages. RTP-LX follows participatory design practices, and a portion of the corpus is especially designed to detect culturally-specific toxic language. We evaluate 10 S/LLMs on their ability to detect toxic content in a culturally-sensitive, multilingual scenario. We find that, although they typically score acceptably in terms of accuracy, they have low agreement with human judges when scoring holistically the toxicity of a prompt; and have difficulty discerning harm in context-dependent scenarios, particularly with subtle-yet-harmful content (e.g. microaggressions, bias). We release this dataset to contribute to further reduce harmful uses of these models and improve their safe deployment.

Taesu Kim, Jongho Lee, Daehyun Ahn et al.

We introduce QUICK, a group of novel optimized CUDA kernels for the efficient inference of quantized Large Language Models (LLMs). QUICK addresses the shared memory bank-conflict problem of state-of-the-art mixed precision matrix multiplication kernels. Our method interleaves the quantized weight matrices of LLMs offline to skip the shared memory write-back after the dequantization. We demonstrate up to 1.91x speedup over existing kernels of AutoAWQ on larger batches and up to 1.94x throughput gain on representative LLM models on various NVIDIA GPU devices.

Taechang Kim, Sooyeon Ji, Kyeongseon Min et al.

$χ$-separation is an advanced quantitative susceptibility mapping (QSM) method that is designed to generate paramagnetic ($χ_{para}$) and diamagnetic ($|χ_{dia}|$) susceptibility maps, reflecting the distribution of iron and myelin in the brain. However, vessels have shown artifacts, interfering with the accurate quantification of iron and myelin in applications. To address this challenge, a new vessel segmentation method for $χ$-separation is developed. The method comprises three steps: 1) Seed generation from $\textit{R}_2^*$ and the product of $χ_{para}$ and $|χ_{dia}|$ maps; 2) Region growing, guided by vessel geometry, creating a vessel mask; 3) Refinement of the vessel mask by excluding non-vessel structures. The performance of the method was compared to conventional vessel segmentation methods both qualitatively and quantitatively. To demonstrate the utility of the method, it was tested in two applications: quantitative evaluation of a neural network-based $χ$-separation reconstruction method ($χ$-sepnet-$\textit{R}_2^*$) and population-averaged region of interest (ROI) analysis. The proposed method demonstrates superior performance to the conventional vessel segmentation methods, effectively excluding the non-vessel structures, achieving the highest Dice score coefficient. For the applications, applying vessel masks report notable improvements for the quantitative evaluation of $χ$-sepnet-$\textit{R}_2^*$ and statistically significant differences in population-averaged ROI analysis. These applications suggest excluding vessels when analyzing the $χ$-separation maps provide more accurate evaluations. The proposed method has the potential to facilitate various applications, offering reliable analysis through the generation of a high-quality vessel mask.

Heejun Shin, Taehee Kim, Jongho Lee et al.

Cone beam computed tomography (CBCT) is an emerging medical imaging technique to visualize the internal anatomical structures of patients. During a CBCT scan, several projection images of different angles or views are collectively utilized to reconstruct a tomographic image. However, reducing the number of projections in a CBCT scan while preserving the quality of a reconstructed image is challenging due to the nature of an ill-posed inverse problem. Recently, a neural attenuation field (NAF) method was proposed by adopting a neural radiance field algorithm as a new way for CBCT reconstruction, demonstrating fast and promising results using only 50 views. However, decreasing the number of projections is still preferable to reduce potential radiation exposure, and a faster reconstruction time is required considering a typical scan time. In this work, we propose a fast and accurate sparse-view CBCT reconstruction (FACT) method to provide better reconstruction quality and faster optimization speed in the minimal number of view acquisitions ($<$ 50 views). In the FACT method, we meta-trained a neural network and a hash-encoder using a few scans (= 15), and a new regularization technique is utilized to reconstruct the details of an anatomical structure. In conclusion, we have shown that the FACT method produced better, and faster reconstruction results over the other conventional algorithms based on CBCT scans of different body parts (chest, head, and abdomen) and CT vendors (Siemens, Phillips, and GE).

Sung Yun Woo, Dongseok Kwon, Byung-Gook Park et al.

Positive-feedback (PF) device and its operation scheme to implement pulse width modulation (PWM) function was proposed and demonstrated, and the device operation mechanism for implementing PWM function was analyzed. By adjusting the amount of the charge stored in the n- floating body (Qn), the potential of the floating body linearly changes with time. When Qn reaches to a threshold value (Qth), the PF device turns on abruptly. From the linear time-varying property of Qn and the gate bias dependency of Qth, fully functionable PWM neuron properties including voltage to pulse width conversion and hard-sigmoid activation function were successfully obtained from a single PF device. A PWM neuron can be implemented by using a single PF device, thus it is beneficial to extremely reduce the area of a PWM neuron circuit than the previously reported one.

Seongbin Oh, Dongseok Kwon, Gyuho Yeom et al.

Hardware-based spiking neural networks (SNNs) are regarded as promising candidates for the cognitive computing system due to low power consumption and highly parallel operation. In this work, we train the SNN in which the firing time carries information using temporal backpropagation. The temporally encoded SNN with 512 hidden neurons showed an accuracy of 96.90% for the MNIST test set. Furthermore, the effect of the device variation on the accuracy in temporally encoded SNN is investigated and compared with that of the rate-encoded network. In a hardware configuration of our SNN, NOR-type analog memory having an asymmetric floating gate is used as a synaptic device. In addition, we propose a neuron circuit including a refractory period generator for temporally encoded SNN. The performance of the 2-layer neural network consisting of synapses and proposed neurons is evaluated through circuit simulation using SPICE. The network with 128 hidden neurons showed an accuracy of 94.9%, a 0.1% reduction compared to that of the system simulation of the MNIST dataset. Finally, the latency and power consumption of each block constituting the temporal network is analyzed and compared with those of the rate-encoded network depending on the total time step. Assuming that the total time step number of the network is 256, the temporal network consumes 15.12 times lower power than the rate-encoded network and can make decisions 5.68 times faster.

Soochang Lee, Chul-Heung Kim, Seongbin Oh et al.

We present a two-layer fully connected neuromorphic system based on a thin-film transistor (TFT)-type NOR flash memory array with multiple postsynaptic (POST) neurons. Unsupervised online learning by spike-timing-dependent plasticity (STDP) on the binary MNIST handwritten datasets is implemented, and its recognition result is determined by measuring firing rate of POST neurons. Using a proposed learning scheme, we investigate the impact of the number of POST neurons in terms of recognition rate. In this neuromorphic system, lateral inhibition function and homeostatic property are exploited for competitive learning of multiple POST neurons. The simulation results demonstrate unsupervised online learning of the full black-and-white MNIST handwritten digits by STDP, which indicates the performance of pattern recognition and classification without preprocessing of input patterns.