• The Journal of the Acoustical Society of Korea
• /
• v.40 no.5
• /
• pp.530-536
• /
• 2021

Recently, due to the development of deep learning, end-to-end speech recognition, which directly maps graphemes to speech signals, shows good performance. Especially, among the end-to-end models, conformer shows the best performance. However end-to-end models only focuses on the probability of which grapheme will appear at the time. The decoding process uses a greedy search or beam search. This decoding method is easily affected by the final probability output by the model. In addition, the end-to-end models cannot use external pronunciation and language information due to structual problem. Therefore, in this paper conformer with lexicon transducer is proposed. We compare phoneme-based model with lexicon transducer and grapheme-based model with beam search. Test set is consist of words that do not appear in training data. The grapheme-based conformer with beam search shows 3.8 % of CER. The phoneme-based conformer with lexicon transducer shows 3.4 % of CER.

• The Journal of the Institute of Internet, Broadcasting and Communication
• /
• v.23 no.5
• /
• pp.151-158
• /
• 2023

The researchers entirely focused on meta-heuristic method for generalized assignment problem(GAP) that is known as NP-hard problem because of the optimal solution within polynomial time algorithm is unknown yet. On the other hand, this paper proposes a heuristic greedy algorithm with rules for finding solutions. Firstly, this paper reduces the weight matrix of original data to w_ij ≤ b_i/l in order to n jobs(items) pack m machines(bins) with l = n/m. The maximum profit of each job was assigned to the machine for the reduced data. Secondly, the allocation was adjusted so that the sum of the weights assigned to each machine did not exceed the machine capacity. Finally, the k-opt swap optimization was performed to maximize the profit. The proposed algorithm is applied to 50 benchmarking data, and the best known solution for about 1/3 data is to solve the problem. The remaining 2/3 data showed comparable results to metaheuristic techniques. Therefore, the proposed algorithm shows the possibility that rules for finding solutions in polynomial time exist for GAP. Experiments demonstrate that it can be a P-problem from an NP-hard.

• The Journal of the Convergence on Culture Technology
• /
• v.6 no.2
• /
• pp.99-106
• /
• 2020

Changam, Lee Samman(1770~1845), who created his own handwriiting to be referred to as the three great writers of the late Joseon Dynasty, invented as the original 'Haengunyusu Typeface' and developed Calligraphy spirit of DonggugJinche in Honam province. He ultimately pursued the state of tonglyeong by raising the personality of 'writing is the person's personality' and the attitude of learning the old. Through the book chang-amseogyeol, he basically polished Haeseo of Han Dynasty and Wi Dynasty and emphasized that Haengseo and Choseo are done automatically when muscle strength and bone strength are established. And since calligraphy originated from 'nature', it goes through the 'Beobcheongwijin' spirit. After doing so, expressed the state of tonglyeong of " mubeob-ibeob ", the stage of reaching. In addition, Changam showed the aesthetic that you can get the novelty by pursuing the philosophy of 'Wu' and the 'beauty of Stupid and Lacking' based on LaoTzu and ChuangTzu. This is a philosophy that follows nature's logic to reveal nature's nature. And it is an aesthetic that protects his 'True Wu' without knowing and greedy. On the other hand, Changam promoted natural and vital beauty through force in the method of using the brush. He suggested the 'Push and Hard' of the Han dynasty, pushing it with force using this power properly. In particular, the feeling of an IlunMujeog brush in 『Changam Calligraphy-The cloud stays Poem』 overflows with the vitality and bizarre and strange dynamism of the spirit and typeface as eum-yang harmonizes with each other. In addition, the beauty of Push and Hard containing polyeoghamse is misaligned, but it has achieved a natural aesthetic without invading. This work demonstrates the real look of Changam choseo. In addition, the beauty of Push and Hard containing polyeoghamse is misaligned, but it has achieved a natural aesthetic without invading. Changam proves the real look of "Haedong's best Chose Maestro".