Abstract: This application relates to a speech synthesis method and apparatus, a model training method and apparatus, and a computer device. The method includes: obtaining to-be-processed linguistic data; encoding the linguistic data, to obtain encoded linguistic data; obtaining an embedded vector for speech feature conversion, the embedded vector being generated according to a residual between synthesized reference speech data and reference speech data that correspond to the same reference linguistic data; and decoding the encoded linguistic data according to the embedded vector, to obtain target synthesized speech data on which the speech feature conversion is performed. The solution provided in this application can prevent quality of a synthesized speech from being affected by a semantic feature in a mel-frequency cepstrum.

Abstract: A method performed by at least one processor includes receiving a first input stream of a task and a second input stream of a solution. The method further includes selecting the first input stream or the second input stream. The method further includes providing the selected input stream to an image conversion model and a language model. The method further includes creating, based on the selected input stream, a model ensemble of the conversion model and the language model. The method further includes outputting a prediction based on the model ensemble. The method may further include generating an image corresponding to text, converting a textual task into a multimodal task, and solving the multimodal task.

Abstract: A method including: receiving an input comprising natural language texts at an encoder; adding a token to the input; obtaining a last-layer hidden state as a natural language text representation; feeding the natural language text representation into a single-layer classification head; predicting a salience allocation based on the single-layer classification head; developing a salience-aware cross-attention (SACA) decoder to determine salience in the natural language text representation; mapping a plurality of salience degrees to a plurality of trainable salience embeddings; estimating an amount of signal to accept from the plurality of trainable salience embeddings; incorporating the salience allocation and the signal in a cross-attention layer model; and generating a summarization based on the SACA decoder and the cross-attention layer model.

Abstract: Provided herein are RNA that: are collectively self-amplifying in an intracellular environment, comprise N1-methyl-pseudouridines and uridines, and have a mole percentage or mole proportion of the N1-methylpseudouridines to the total of the uridines and the Ni-methylpseudouridines or a mole ration of the N1-methylpseudouridines to the uridines.

Abstract: A method for zero-shot event extraction, performed by a computer device. The method includes training a context encoder, a first definition encoder, and a second definition encoder with auto extracted context-definition alignment data; retrieving a plurality of verbal synsets from a lexical database; refining a representation model based on the context-definition alignment data and the plurality of verbal synsets; encoding a plurality of candidate event type definitions; encoding the refined representation model with the trained context encoder; and determining whether the encoded representation model belongs to one of the plurality of candidate event type definitions based on a cosine similarity between the encoded representation model, the trained context encoder, the first trained definition encoder, and the second trained definition encoder.

Abstract: A method, computer program, and computer system is provided for identifying a speaker in at text based work. Labeled and unlabeled instances corresponding to one or more speakers are extracted. Pseudo-labels are inferred for the extracted unlabeled instances based on the labeled instances. One or more of the unlabeled instances are labeled based on the inferred pseudo-labels.

Abstract: A mixed speech recognition method, a mixed speech recognition apparatus, and a computer-readable storage medium are provided. The mixed speech recognition method includes: monitoring an input of speech input and detecting an enrollment speech and a mixed speech; acquiring speech features of a target speaker based on the enrollment speech; and determining speech belonging to the target speaker in the mixed speech based on the speech features of the target speaker. The enrollment speech includes preset speech information, and the mixed speech is non-enrollment speech inputted after the enrollment speech.

Abstract: An electronic device includes a first part and a second part that can be folded or expanded relative to each other. The electronic device has a closed state and an open state When the electronic device is in the closed state, a frame of the first part and a frame of the second part partially or totally overlap. The first part includes a first feeding antenna. The second part includes a first parasitic antenna, and when the electronic device is in the closed state, the first parasitic antenna is not grounded and can be coupled to the first feeding antenna.

Abstract: Methods and apparatuses are provided for performing sequence to sequence (Seq2Seq) speech recognition training performed by at least one processor. The method includes acquiring a training set comprising a plurality of pairs of input data and target data corresponding to the input data, encoding the input data into a sequence of hidden states, performing a connectionist temporal classification (CTC) model training based on the sequence of hidden states, performing an attention model training based on the sequence of hidden states, and decoding the sequence of hidden states to generate target labels by independently performing the CTC model training and the attention model training.

Abstract: An unsupervised text to speech system utilizing a lexicon to map input text to the phoneme sequence, which is expanded to the frame-level forced alignment with a speaker-dependent duration model. An alignment mapping module that converts the forced alignment to the unsupervised alignment (UA). Afterword, a Conditional Disentangled Sequential Variational Auto-encoder (C-DSVAE), serving as the self-supervised TTS AM, takes the predicted UA and a target speaker embedding to generate the mel spectrogram, which is ultimately converted to waveform with a neural vocoder.

Abstract: An antenna apparatus used in an electronic device having a flexible display, the flexible display can be bent at a rotating shaft, the flexible display includes a primary screen and a secondary screen respectively configured on two sides of the rotating shaft. The antenna apparatus may include a first metal strip disposed on the primary screen frame close to one end of the rotating shaft, and a second metal strip disposed on the secondary screen frame close to the same end of the rotating shaft. The first metal strip may be implemented as a plurality of antennas through dual-feed design. When the flexible display is in a folded state, the second metal strip may be coupled to the first metal strip to generate radiation. In this case, the second metal strip may be used as a parasitic antenna of the first metal strip.

Abstract: The present disclosure relates to a composite logic gate circuit, including: a simple logic gate circuit including a first logic gate circuit and an inverter circuit, a first PMOS transistor, and a first NMOS transistor. The first logic gate circuit is configured to receive a first input signal and a second input signal, and to output a first output signal. The inverter circuit includes a second PMOS transistor and a second NMOS transistor. A source of the second PMOS transistor is coupled to a power input terminal, a drain is coupled to a drain of the second NMOS transistor, and a gate is configured to receive the first output signal. A source of the second NMOS transistor is coupled to a ground terminal, and a gate is configured to receive the first output signal. A source of the first PMOS transistor is coupled to the drain of the second PMOS transistor, a drain is coupled to a drain of the first NMOS transistor, and a gate is configured to receive a third input signal.

Abstract: An image outputting device includes a sensing circuit for generating an image signal according to a configuration; a processing circuit, coupled to the sensing circuit, for performing an image processing on the image signal according to the configuration to generate an image processing result; and a controlling circuit, coupled to the sensing circuit and the processing circuit, for setting the configuration and entering an operating system after setting the configuration.

Abstract: An antenna includes a patch radiator, a feed point, a first ground point, and a second ground point. The patch radiator has a first side edge and an intersecting second side edge, and a first coupling contact and a second coupling contact. The patch radiator is coupled to the first ground point and the second ground point through the first coupling contact and the second coupling contact, and is biased coupled to the feed point. A distance between the first coupling contact and the first side edge, a distance between the first coupling contact and the second side edge, a distance between the second coupling contact and the first side edge, and a distance between the second coupling contact and the second side edge are all greater than or equal to 0.05?, where ? is an operating wavelength of the antenna in an operating frequency band range of the antenna.

Abstract: A method of acoustic echo suppression using a recurrent neural network, performed by at least one processor, is provided. The method includes receiving a microphone signal and a far-end reference signal, estimating an echo suppressed signal and an echo signal based on the microphone signal and the far-end reference signal, estimating enhancement filters for the microphone signal based on the echo suppressed signal and the echo signal, generating an enhanced signal based on the enhancement filters, and adjusting the enhanced signal using automatic gain control (AGC) and outputting the adjusted signal.

Abstract: Methods and devices to efficiently normalize text by processing inputted text based on a text normalization model that includes processing the input text in a first stage including a statistical model as a first output, processing the first output in a second stage including a rule based model as a normalized text, and outputting the normalized text.

Abstract: A spectator interaction method includes displaying an interactive control for a first spectator account watching an online battle, the interactive control configured to trigger an interaction between the first spectator account and one or more spectator accounts currently watching the online battle. The method further includes displaying first interactive content in response to receiving a trigger operation on the interactive control. The method further includes displaying a quantity of spectator accounts corresponding to the first interactive content, the quantity of spectator accounts including accounts that have interacted with the first interactive content, and accounts that have interacted with second interactive content triggered by a second spectator account. The second interactive content being a same interactive content type as the first interactive content.

Abstract: An antenna apparatus includes a grounding plate, a monopole, a first feeding component, and a second feeding component. A slot on the grounding plate includes a first slot and a second slot that interpenetrate each other, the second slot extends from the first slot to an edge of the grounding plate. The monopole includes a first stub and a second stub extending from the first stub to the second slot, the second stub and the second slot form a feeding structure. The first feeding component is electrically coupled to the grounding plate to feed the feeding structure to excite a first radiation mode of the antenna apparatus. The second feeding component is electrically coupled to the second stub to feed the feeding structure to excite a second radiation mode of the antenna apparatus. Polarizations in the two radiation modes are orthogonal.

Abstract: The present disclosure provides a multi-state one-time programmable (MSOTP) memory circuit including a memory cell and a programming voltage driving circuit. The memory cell includes a MOS storage transistor, a first MOS access transistor and a second MOS access transistor electrically connected to store two bits of data. When the memory cell is in a writing state, the programming voltage driving circuit outputs a writing control potential to the gate of the MOS storage transistor, and when the memory cell is in a reading state, the programming voltage driving circuit outputs a reading control potential to the gate of the MOS storage transistor.

Abstract: The present invention relates to a groundwater circulation well system with pressure-adjustable hydrodynamic cavitation, including a circulation well body, a sucked and injected water circulation assembly and a hydrodynamic cavitator. The sucked and injected water circulation assembly is based on a water suction and injection pump. The hydrodynamic cavitator is provided, inside a vortex chamber, with a vortex water inlet column capable of changing a water passing aperture. The hydrodynamic cavitator is capable of changing a bubbling pressure and a breaking pressure of hydrodynamic cavitation bubbles in the vortex water inlet column. The hydrodynamic cavitator generates vortices in the circulation well body to accelerate uniform mixing of a remediation agent and the groundwater. Energy from collapsing and bursting of the hydrodynamic cavitation bubbles activates the remediation agent such that contaminants in the groundwater are efficiently degraded.