Abstract: A method in an illustrative embodiment comprises collecting a plurality of audio signals from a plurality of sensors, and combining the plurality of audio signals to generate synthetic audio, wherein the plurality of sensors are located at different locations in the same space, and the plurality of audio signals each have a location identifier. The method further comprises analyzing the synthetic audio to find an audio subset in the plurality of audio signals, and extracting first image features of a first image of the space. The method further comprises modifying the first image features based on the location identifiers of the audio signals in the audio subset to generate second image features, and generating a second image of the space based on the second image features.

Abstract: Embodiments of the present disclosure relate to a method, an electronic device, and a computer program product for generating a three-dimensional image. The method includes: receiving a first image presenting a target object at a first viewing angle, wherein the first image is a two-dimensional image; determining a transformed image of the first image at a target viewing angle, wherein the target viewing angle is the same as or different from the first viewing angle. The method further includes: generating a first representation using a first feature extraction layer corresponding to the first viewing angle in an encoder based on the transformed image; and generating a second image based on the first representation, wherein the second image is a three-dimensional image and presents the target object at the target viewing angle.

Abstract: Embodiments of the present disclosure relate to a method, a device, and a computer program product for determining a node of a decision tree. The method includes determining multiple features of multiple modals corresponding to input information. The method further includes generating a multi-modal feature representation by combining the multiple features of the multiple modals. The method further includes determining a target path in a decision tree that is associated with the multi-modal feature representation, the decision tree comprising multiple nodes. The method further includes determining, in the target path based on the multi-modal feature representation, a target node associated with the input information and used to indicate a question or an answer. This method enables the fusion of feature representations corresponding to input information of different modals to determine a multi-modal feature representation. In this way, it is possible to determine richer and more accurate user intentions.

Abstract: Methods, apparatus, and processor-readable storage media for artificial intelligence-driven avatar-based personalized learning techniques are provided herein.

Abstract: Example embodiments of the present disclosure provide a method, a device, and a computer program product for data query. The method includes selecting, according to a type of input data, a target pre-trained model from a deep network pool including a plurality of pre-trained models; performing, by using the selected target pre-trained model, feature extraction on the input data to determine text descriptors for the input data; and generating, based on the text descriptors, a query table for query. The method according to the present disclosure can select, according to different input data, different target pre-trained models from the deep network pool including the plurality of pre-trained models to process (e.g., compress) the input data. The method according to the present disclosure assembles a plurality of deep networks into a pool to automatically process data to obtain text descriptors for data retrieval, thereby achieving efficient data compression and retrieval.

Abstract: A method in an illustrative embodiment includes acquiring a first sequence of a set of image frames of a video arranged in positive order. The method further includes generating, based on the first sequence of a positive-order arrangement, a second sequence of the set of image frames arranged in reverse order. The method further includes determining a first cross correlation sequence of the image frames in the first sequence relative to the image frames in the second sequence; determining a second cross correlation sequence of the image frames in the second sequence relative to the image frames in the first sequence; determining, based on the first cross correlation sequence and the second cross correlation sequence, a global feature distribution for the set of image frames; and finally, determining, based on the global feature distribution, at least one target frame in the set of image frames.

Abstract: Embodiments of the present disclosure relate to a method, a device, and a computer program product for video retrieval. The method includes determining a retrieval level corresponding to a retrieval word in response to receiving a retrieval request including the retrieval word from a client. The method further includes determining a video database corresponding to the retrieval level among a plurality of video databases, where the plurality of video databases store image frames for different frame rates of the same video. The method further includes retrieving an image frame associated with the retrieval word from the determined video database and sending the retrieved image frame to the client. According to the solution, multi-level video retrieval can be realized, allowing a user to retrieve a desired video in different scenes or different devices at different retrieval speeds, so as to provide the user with a more flexible video retrieval mode.

Abstract: Embodiments of the present disclosure relate to a method, a device, and a computer program product for data augmentation. The method includes generating an image embedding based on an image in an unstructured document, and generating a text embedding based on text in the unstructured document and associated with the image. The method further includes acquiring descriptive information from a storage library based on the generated image embedding and text embedding. The method further includes adding the acquired descriptive information into the unstructured document. In this way, it can be possible not only to understand and analyze the unstructured document across modalities, but also to enrich it with a characterization of multimodal data in the unstructured document, thus increasing the amount and diversity of data.

Abstract: Embodiments of the present disclosure relate to a method, an electronic device, and a computer program product for a chatbot. The method includes determining, based on a query entered by a user to a chatbot, a first representation associated with the query. The method further includes generating, based on the first representation and a domain to which the query belongs, a second representation, wherein dimensions of the second representation are smaller than those of the first representation. The method further includes generating, by a decoder corresponding to the domain based on the second representation, a response to the query. With embodiments of the present disclosure, quality of the generated response to the query and consistency of the response can be improved, and universality and specificity of the response can be balanced.

Abstract: Embodiments of the present disclosure relate to a method, an electronic device, and a computer program product for constructing training data. The method includes determining multiple clusters by clustering prompts in a training dataset; and determining, based on multiple cohesion levels of the multiple clusters, multiple sampling probabilities corresponding to the multiple clusters, where the cohesion levels indicate intra-cluster distances in the clusters. The method further includes determining, according to the multiple sampling probabilities, a target cluster for sampling. The method further includes constructing target training data by sampling target prompts from the target cluster.

Abstract: Example embodiments of the present disclosure provide a method, a device, and a computer program product for processing a workflow chart. The method includes encoding structural information of the workflow chart including a plurality of nodes and a plurality of edges by using a graph neural network to acquire a vector representation of the structural information; acquiring textual description data about the workflow chart at the nodes; training a language model based on the acquired textual description data and the acquired vector representation to acquire a pretrained language model; and fine-tuning the pretrained language model through training data of a specific task to acquire a fine-tuned language model. Through the method for processing the workflow chart of the present disclosure, the combination of the graph neural network and the language model not only can process a large number of complex workflow charts, but also can generate effective natural language outputs.

Abstract: Embodiments of the present disclosure relate to a method, an electronic device, and a computer program product for generating an image. The method includes acquiring a semantic segmentation graph by performing semantic segmentation on a source image. The method further includes acquiring a key word for describing a feature of a to-be-generated target image. The method further includes transforming the semantic segmentation graph by using the key word so as to acquire a transformed semantic segmentation graph. The method further includes generating the target image based on the transformed semantic segmentation graph. According to the method of embodiments of the present disclosure, a semantic segmentation graph of a source image and a key word can be used to generate a target image, so as to make the generated target image have a target feature and have semantic consistency with the source image, thereby generating a high-quality target image.

Abstract: In a method for managing training data in an illustrative embodiment, in response to a determination that new training data is collected by a sensor, the new training data is stored into a collected data stream of a storage pool; in response to a determination that the new training data and historical data stored in a full data stream of the storage pool are refined into refined training data, the refined training data is stored into a refined data stream of the storage pool; and the new training data is stored into the full data stream. In this way, data streams become clear and storage costs are reduced. This greatly reduces the difficulty of developing complex scenarios such as autonomous driving.

Abstract: Techniques for constructing and otherwise managing knowledge graphs in information processing system environments are disclosed. For example, a method comprises the following steps. The method collects data from a plurality of data sources. The method extracts structured data and unstructured data from the collected data, wherein unstructured data is extracted using an unsupervised machine learning process. The method forms a plurality of sub-graph structures comprising a sub-graph structure for each of the data sources based on at least a portion of the extracted structured data and unstructured data. The method combines the plurality of sub-graph structures to form a combined graph structure representing the collected data from the plurality of data sources. The resulting combined graph structure is a comprehensive knowledge graph.

Abstract: The present disclosure relates to a method, a device, and a product for searching data. The method includes: encoding a search input into a first dense vector based on a first multi-modal search model; determining, based on the first dense vector, a distilled data item corresponding to the search input from a distilled dataset corresponding to the first multi-modal search model; encoding, based on the first multi-modal search model, an original data item in an original data subset corresponding to the distilled data item in an original dataset corresponding to the distilled dataset into a second dense vector; and determining, based on the second dense vector, an original data item from the original data subset as a search result corresponding to the search input. The method for searching data according to the present disclosure can improve the efficiency and security of data storage, model reproduction, and multi-modal data management.

Abstract: Embodiments of the present disclosure relate to a method for generating a machine learning model. The method includes: encoding a decision tree using a graph neural network; inputting the encoded decision tree into a machine learning model; generating a question corresponding to each node of a root node and internal nodes of the decision tree using the machine learning model; inputting a natural language text to the machine learning model; and generating the machine learning model using the inputted natural language text and the generated question. According to embodiments of the present disclosure, a machine learning model that enables faster and more accurate provision of corresponding solutions based on the inputted natural language text can be realized.

Abstract: Embodiments of the present disclosure relate to a method, an electronic device, and a computer program product for content recommendation. A method for content recommendation includes determining a similarity between a first recommendation result and a second recommendation result for a content set. The first recommendation result and the second recommendation result are determined based on different recommendation techniques and respectively indicative of a recommendation degree for each content in the content set. The method further includes adjusting the second recommendation result using the similarity. In addition, the method further includes determining a target recommendation result for the content set based on the first recommendation result and the adjusted second recommendation result. In this manner, the accuracy and stability of the finally obtained recommendation results can be improved.

Abstract: The present disclosure relates to a method, a device, and a computer program product for generating video. The method includes extracting a first text of a first language in an input video and generating a second text of a second language corresponding to the first text of the first language. The method further includes generating a gist frame of the input video based on the input video and the second text, where the gist frame indicates information associated with color configuration and object layout of the input video. In addition, the method further includes generating, based on the gist frame and the second text, an output video broadcasting the second text by means of the second language. The method of the present disclosure enhances the similarity of the output video and the input video, and expressions and actions of a narrator can be generated at a correct time point.

Abstract: In an illustrative embodiment, a method is disclosed for generating an image from fMRI data and EEG data. The method includes extracting a first feature map of the fMRI data, the first feature map being multidimensional data having a temporal dimension related to a sample collection time of the fMRI data, and converting the first feature map based on the temporal dimension of the first feature map and a first attention model. The method further includes extracting a second feature map of the EEG data, the second feature map being multidimensional data having a spatial dimension related to an electrode for collecting the EEG data, and converting the second feature map based on the spatial dimension of the second feature map and a second attention model. The method further includes generating an image based on the converted first feature map and the converted second feature map.

Abstract: Embodiments of the present disclosure relate to a method for generating a machine learning model. The method includes extracting multiple parameters from a target machine learning model, where the multiple parameters include a learning rate, state information, a loss value, a gradient, and a weight, and the target machine learning model is configured to execute tasks related to at least one of images, videos, voice, and text. The method further includes predicting a first learning rate by a first machine learning model based on the multiple parameters; predicting a second learning rate by a second machine learning model based on the multiple parameters; choosing, based on the first learning rate and the second learning rate, a learning rate having a minimum loss value in the first learning rate and the second learning rate; and adjusting the target machine learning model based on the learning rate having the minimum loss value.