Search

This paper presents a method for building a personalized open-domain dialogue system to address the WWH (WHAT, WHEN, and HOW) problem for natural response generation in a commercial setting, where personalized dialogue responses are heavily interleaved with casual response turns. The proposed approach involves weighted dataset blending, negative persona information augmentation methods, and the design of personalized conversation datasets to address the challenges of WWH in personalized, open-domain dialogue systems. Our work effectively balances dialogue fluency and tendency to ground, while also introducing a response-type label to improve the controllability and explainability of the grounded responses. The combination of these methods leads to more fluent conversations, as evidenced by subjective human evaluations as well as objective evaluations., Accepted in ACL 2023 Industry Track

Multi-turn response selection models have recently shown comparable performance to humans in several benchmark datasets. However, in the real environment, these models often have weaknesses, such as making incorrect predictions based heavily on superficial patterns without a comprehensive understanding of the context. For example, these models often give a high score to the wrong response candidate containing several keywords related to the context but using the inconsistent tense. In this study, we analyze the weaknesses of the open-domain Korean Multi-turn response selection models and publish an adversarial dataset to evaluate these weaknesses. We also suggest a strategy to build a robust model in this adversarial environment., EMNLP 2021

The metabolic activity of microorganisms is considered one of the reasons why evaporators of the automobile and household air conditioning systems produce the malodor. The odor-causing by microorganisms can be reduced by incorporating biocide in the hydrophilic coating. The objective of this work was to understand the mechanisms how the microorganisms produced the odor and to analyze the chemical compound of the odor. Experiments were conducted to measure the variation of odor and to assess the antimicrobial property of epoxy resin-based hydrophilic coatings of evaporators. Then the microorganisms used the experiments were as followed; four bacteria cultured from an evaporator of household air conditioning system were metylobacterium phylloshaerae, metylobacterium radiotolerans, sphingomonas aquatilis, and bacillus subtilis, and three molds were aspergillus niger (ATCC 6275), paecilomyces variotii (ATCC 18502), and trichoderma viren (ATCC 9645). Then the four bacteria and three fungi mixed with similar concentration respectively and were recoverable repeatedly for seven days. The change of intensity and types of odor-causing chemicals as a result of their metabolic activity were measured by Gas chromatography-mass spectrometry (GC-MS) measuring and sensory evaluation of odor. In order to compare the number of microorganisms, the samples were incubated for two days at 35°C and R2A agar was employed for the isolation of bacteria and fungi.

We experimentally investigate the frosting and defrosting performance of slippery surfaces which have low sliding angles of water droplets and low ice adhesion strengths. The frosting and defrosting characteristics of slippery surfaces are compared with those of bare aluminum, hydrophilic, and superhydrophobic surfaces. The enhanced sliding properties of water droplets on the slippery surfaces effectively promote the drainage of the condensate on their surfaces, which not only leads to significant frost retardation under frosting conditions, but also substantially reduces the mass of the retained water on the surfaces after defrosting by heat. In addition, when mechanical vibration is applied together with heating during defrosting process, the low ice adhesion strengths of the slippery surfaces enable the effective detachment of the lumped frost layer from their surfaces, thereby significantly reducing the defrosting time.