Long Short-Term Memory Recurrent Neural Network-Based Acoustic Model Using Connectionist Temporal Classification on a Large-Scale Training Corpus

A Long Short-Term Memory(LSTM) Recurrent Neural Network(RNN) has driven tremendous improvements on an acoustic model based on Gaussian Mixture Model(GMM). However, these models based on a hybrid method require a forced aligned Hidden Markov Model(HMM) state sequence obtained from the GMM-based acoustic model. Therefore, it requires a long computation time for training both the GMM-based acoustic model and a deep learning-based acoustic model. In order to solve this problem, an acoustic model using CTC algorithm is proposed. CTC algorithm does not require the GMM-based acoustic model because it does not use the forced aligned HMM state sequence. However, previous works on a LSTM RNN-based acoustic model using CTC used a small-scale training corpus. In this paper, the LSTM RNN-based acoustic model using CTC is trained on a large-scale training corpus and its performance is evaluated. The implemented acoustic model has a performance of 6.18% and 15.01% in terms of Word Error Rate(WER) for clean speech and noisy speech, respectively. This is similar to a performance of the acoustic model based on the hybrid method.

Nondestructive Skeletal Imaging of Hyla suweonensis Using Micro-Computed Tomography

We successfully obtained 3D skeletal images of Hyla suweonensis, employing a nondestructive method by applying appropriate anesthesia and limiting the radiation dose. H. suweonensis is a tree frog endemic to Korea and is on the list of endangered species. Previous studies have employed caliper-based measurements and two-dimensional （2D） X-ray imaging for anatomical analyses of the skeletal system or bone types of H. suweonensis. In this work we reconstructed three-dimensional （3D） skeletal images of H. suweonensis, utilizing a nondestructive micro-computed tomography （micro-CT） with a short scan and low radiation dose （i.e. 4 min and 0.16 Gy）. Importantly, our approach can be applied to the imaging of 3D skeletal systems of other endangered frog species, allowing both versatile and high contrast images of anatomical structures without causing any significant damages to the living animal.

Language Model Using Differentiable Neural Computer Based on Forget Gate-Based Memory Deallocation

A differentiable neural computer(DNC)is analogous to the Von Neumann machine with a neural network controller that interacts with an external memory through an attention mechanism.Such DNC’s offer a generalized method for task-specific deep learning models and have demonstrated reliability with reasoning problems.In this study,we apply a DNC to a language model(LM)task.The LM task is one of the reasoning problems,because it can predict the next word using the previous word sequence.However,memory deallocation is a problem in DNCs as some information unrelated to the input sequence is not allocated and remains in the external memory,which degrades performance.Therefore,we propose a forget gatebased memory deallocation(FMD)method,which searches for the minimum value of elements in a forget gate-based retention vector.The forget gatebased retention vector indicates the retention degree of information stored in each external memory address.In experiments,we applied our proposed NTM architecture to LM tasks as a task-specific example and to rescoring for speech recognition as a general-purpose example.For LM tasks,we evaluated DNC using the Penn Treebank and enwik8 LM tasks.Although it does not yield SOTA results in LM tasks,the FMD method exhibits relatively improved performance compared with DNC in terms of bits-per-character.For the speech recognition rescoring tasks,FMD again showed a relative improvement using the LibriSpeech data in terms of word error rate.

Overcoming the obstacles of current photodynamic therapy in tumors using nanoparticles

Photodynamic therapy(PDT)has been applied in clinical treatment of tumors for a long time.However,insufficient supply of pivotal factors including photosensitizer(PS),light,and oxygen in tumor tissue dramatically reduces the therapeutic efficacy of PDT.Nanoparticles have received an influx of attention as drug carriers,and recent studies have demonstrated their promising potential to overcome the obstacles of PDT in tumor tissue.Physicochemical optimization for passive targeting,ligand modification for active targeting,and stimuli-responsive release achieved efficient delivery of PS to tumor tissue.Various trials using upconversion NPs,two-photon lasers,X-rays,and bioluminescence have provided clues for efficient methods of light delivery to deep tissue.Attempts have been made to overcome unfavorable tumor microenvironments via artificial oxygen generation,Fenton reaction,and combination with other chemical drugs.In this review,we introduce these creative approaches to addressing the hurdles facing PDT in tumors.In particular,the studies that have been validated in animal experiments are preferred in this review over proof-of-concept studies that were only performed in cells.