316 results about "Sequence processing" patented technology

An example method can include providing a natural language instruction and user interface image data to a machine-learned sequence processing model that is configured to process image data and generate commands for controlling the target computing device, wherein the machine-learned sequence processing model has parameters learned using an interface recognition objective based on an evaluation of an interface recognition output generated based on processing a rendered training interface from a pre-training dataset and an interface navigation objective based on an evaluation of a user interface command generated based on processing a rendered training interface from a fine-tuning dataset; receiving, from the machine-learned sequence processing model, a command indicating an interaction with the user interface to implement the natural language instruction; and generating, based on the command, a control signal configured to initiate the interaction.

Provided is a methodology for direct supervised preference fine-tuning of sequence processing models such as, for example, so-called large language models (LLMs) and large multimodal models (LMMs). The proposed approaches can fine-tune the model to directly predict the posterior token probabilities conditioned on a positive preference of the sequence for which the token is the last token on a sequence of tokens that are the prefix to the sequence. This method offers a simpler fine-tuning approach that directly generates the desired posteriors for use in decoding, without requiring additional inference per vocabulary token at decoding time.

An example method includes receiving input requests to process a plurality of input sequences using the machine-learned sequence processing model to generate a plurality of output sequences respectively corresponding to the plurality of input sequences; generating a plurality of initial attention tensors respectively for the plurality of input sequences, wherein: one or more respective initial attention tensors are generated for each respective input sequence in parallel over input elements of the respective input sequence; and the one or more respective initial attention tensors are generated in one or more batches having a first batch size using a prefill system that comprises one or more prefill computing devices and executes one or more layers of the machine-learned sequence processing model; and autoregressively generating, using the plurality of initial attention tensors, a plurality of output elements for each of the plurality of output sequences in one or more batches having a second batch size, wherein: the plurality of output elements are generated using a generation system that comprises one or more generation computing devices.

The invention discloses a metering error correction method and system based on electric power data acquisition, and the method comprises the steps: detecting an operation state through a self-inspection and data acquisition module, and obtaining environment data, electrical parameters and error information; key feature data are selected and standardized through a data preprocessing module; an LSTM model and an SVR model are constructed through a model establishment and fusion module, and prediction results are fused to generate error compensation output; real-time error compensation is carried out on a combined prediction result by utilizing an ARIMA model and error time sequence processing through a dynamic error compensation module; the data smoothing and evaluation module is used for smoothing the voltage data and evaluating the error compensation effect; an error compensation effect is monitored through a monitoring and optimizing module, and feature selection and model parameters are optimized regularly; through the fault diagnosis and emergency module, an abnormal condition is detected, and a corresponding standby scheme is triggered. According to the invention, the accuracy and anti-interference capability of electric power metering are effectively improved, and the method is suitable for complex and changeable metering environments.

Provided is a generative sequence processing model that is specifically finetuned for cybersecurity applications. This specialized generative sequence processing model can be finetuned on a comprehensive range of cybersecurity data and associated finetuning tasks, providing the model with rich capabilities for analyzing, understanding, describing, and taking action with respect to the real-time cybersecurity data generated by one or more cybersecurity operations tools. Thus, the creation and use of the generative sequence processing model represents a solution to the technical challenge of usefully analyzing and acting upon a large volume of data generated by a number of disparate cybersecurity operations tools deployed by an organization.

The invention provides an image restoration and super-resolution reconstruction system and method based on deep learning, and belongs to the technical field of digital image processing. The invention aims to solve the problems of high calculation complexity and resource consumption, limitation of long sequence processing, high training difficulty and texture scene deficiency when a multi-scale residual network based on a Transform architecture is used for image resolution conversion. The reconstruction system comprises: an image preprocessing module performing window division and video memory optimization on an input low-resolution image; the multi-layer fusion network dynamically adjusts the characteristics of the low-resolution image, captures channel information in different scenes, performs interactive fusion, performs comparison supervision, establishes an information communication channel, dynamically adjusts and optimizes parameters through negative feedback, and obtains a super-resolution image. And the loss function module maximizes the similarity of the super-resolution image and the high-resolution image in the segmentation feature space to obtain a final super-resolution image.

Aspects of the disclosed technology include computer-implemented systems and methods for updating machine-learned models while preserving adaptations of the models by downstream applications. The system can obtain data indicative of a respective adapter implemented by each of a plurality of downstream applications for a first version of a machine-learned sequence processing model. The system can generate a first training loss based on processing a set of training data by a second version of the machine-learned sequence processing model. The system can generate a second training loss based on processing the set of training data using the respective adapter implemented by each of the plurality of downstream applications. The system can modify at least a portion of the second version of the core sequence processing model using the first training loss and the second training loss.

Provided is a framework for fine-tuning pre-trained sequence processing models to human preferences and / or other objective(s). Instead of using reinforcement learning to fine-tune the LLM parameters towards the human preferences, example systems take a Bayesian approach which can preserve the learned prediction distributions of the pre-trained model, but adds explicit sequential preference tuned predictions in a multi-objective model fine-tuning training setup. The model can be tuned to predict posterior token probabilities conditioned on the human preferences.

The invention discloses a big language model-oriented inference operation acceleration method for processing an extremely long text, which belongs to the technical field of large-scale pre-training language model inference optimization, and specifically comprises the following steps of: selecting a big language model, starting a sparse attention mode, inputting different texts, recording execution time of each attention head and an attention mode type in different layers, and calculating the inference operation speed of the large language model; establishing a statistical database; solving a multi-head attention load balanced distribution strategy of the large language model under the actual text; establishing a weight index table by splitting the weight matrix; retrieving a weight sub-matrix corresponding to each attention head, and loading the weight sub-matrix to a corresponding GPU device; by asynchronously preloading the MHA calculation weight and the MLP calculation weight of the adjacent layers and combining KV cache management, load balancing in the reasoning process is realized. According to the method, dynamic segmentation, recombination and scheduling of the attention head are carried out in the reasoning stage, so that the problems of multi-GPU load imbalance and resource idling are effectively avoided, and the system throughput of long sequence processing is remarkably improved.

The invention belongs to the related technical field of artificial intelligence, and discloses a robot time sequence imitation learning method and system based on a Mama coding complete history, and the method comprises the steps: receiving a multi-modal observation sequence in a task execution process of a robot, the multi-modal observation sequence comprising observation data of at least one sensor; processing the multi-modal observation sequence by using a sequence processing module based on a state space model, and updating a time sequence output of complete historical information of one code up to the current time step at each time step; and predicting the next step or a series of future actions of the robot based on the time sequence output of the current time step so as to control the robot to simulate. The time sequence processing module based on the state space model is utilized to process and encode the complete observation history in the task execution process of the robot, so that a non-Markov decision-making imitation learning method is realized, and the learning efficiency and the execution success rate of the robot in a complex and state-dependent long time sequence operation task are improved.

An example method is provided for training a machine-learned student sequence processing model, the method comprising: obtaining a respective input; obtaining, from the student machine-learned sequence processing model, a respective output corresponding to the respective input; generating a multiscale refinement objective configured to jointly distill knowledge from a teacher machine-learned sequence processing model and reinforce preferred behavior of the student machine-learned sequence processing model, wherein the multiscale refinement objective comprises: a first component based on a divergence metric characterizing, for the respective input, a comparison of a plurality of predictions of the student machine-learned sequence processing model to a plurality of predictions of the teacher machine-learned sequence processing model; and a second component based on a reinforcement learning signal associated with the respective output; and updating the machine-learned student sequence processing model based on the multiscale refinement objective.

The invention discloses a Token-based visual task generation method, and belongs to the technical field of intelligent task automation, and the method comprises the steps: S1, cross-modal alignment; s2, performing visual Token processing; s3, constructing a task description Token sequence: constructing the task description Token sequence based on a predefined visual task template library according to requirements of a user or a specific application scene; s4, checking task feasibility; s5, task priority scheduling; s6, training a task generation model; s7, dynamic task allocation; and S8, optimizing the model. According to the method, the long sequence processing capability is optimized through a hierarchical merging strategy, the compactness of feature expression is realized while space position information is reserved, linear projection and enhanced position coding are combined to form a visual Token sequence with strong representation capability, local detail features are contained, a global context relationship is kept, and the method is suitable for the visual Token sequence with high representation capability. High-information-density feature input is provided for subsequent task processing, and the processing precision of various visual algorithms is effectively improved.

An artificial intelligence-based image processing system comprises a processor that executes instructions stored on a memory to classify an input image with a prototypical part neural network including a backbone subnetwork, a prototype subnetwork, and a readout subnetwork to produce an interpretable classification of the input image including one or a combination of a classification result of the input image and an interpretation of the classification result. The backbone subnetwork is trained with machine learning to process the input image with an incomplete sequence of active convolutional layers producing feature embeddings representing features extracted from pixels of different regions of the input image. The prototype subnetwork is trained to compare the feature embeddings with prototypical feature embeddings to produce results of comparison and the readout subnetwork is configured to analyze the results of comparison to produce the interpretable classification of the input image.

A machine-learned system for aligning textual and image representations prior to input to a sequence processing model is described. The system includes a machine-learned image embedding model configured to receive image data and generate one or more image embeddings and a machine-learned text embedding model configured to receive text data and the one or more image embeddings and generate one or more text embeddings. The system includes a machine-learned cross-modal adapter configured to generate one or more text tokens aligned with one or more image tokens based at least in part on aligning data associated with the one or more text embeddings and the one or more image tokens. The system includes a machine-learned sequence processing model configured to generate an output based at least in part on the one or more text tokens and the one more image tokens.

The present disclosure provides systems and methods that align sequence processing models with recommendation knowledge. Example training systems can generate natural language prompts, which can be referred to as ‘auxiliary prompts’, that encode different types of recommendation-related knowledge, such as item attributes and user preferences. These auxiliary prompts encode into natural language format various operations and losses that can be used to impart recommendation knowledge to a sequence processing model, including item embedding, Bayesian personalized ranking (BPR), and masked item modeling.

A computing device can obtain a first machine-learned sequence processing model configured to use a plurality of first tools, wherein at least one first tool of the plurality of first tools is a second machine-learned sequence processing model configured to use one or more second tools. The computing device can obtain an input context. The computing device can select, using the first machine-learned sequence processing model based at least in part on the input context, a first tool of the plurality of first tools, wherein the first tool selected is the second machine-learned sequence processing model. The computing device can select, using the second machine-learned sequence processing model, at least one second tool of the one or more second tools. The computing device can generate, using the at least one second tool of the one or more second tools, a first output.

Systems and methods are provided for efficiently calibrating a data mixture for training machine-learned models (e.g., machine-learned sequence processing models, such as transformer-based models). For example, machine-learned models can be trained over a broad dataset that can include multiple different categories of data. The mixture of data categories within the dataset can influence model performance. To improve the performance of machine-learned models, example implementations of the present disclosure can learn a distribution of data categories using a lightweight proxy model before initiating training of a large primary model. In this manner, for instance, example implementations can obtain an improved training data distribution with less computational expense and can leverage the learned training data distribution to better train a large primary model.

The invention belongs to the field of artificial intelligence and electroencephalogram emotion recognition, and provides a double-branch electroencephalogram emotion recognition method and system based on brain region topology and time-space, and the method comprises the steps: preprocessing a to-be-recognized electroencephalogram signal to obtain a plurality of electroencephalogram fragments, and extracting a difference entropy sequence of each electroencephalogram fragment and a Spearman correlation coefficient matrix between channels; based on the Spearman correlation coefficient matrix, utilizing a bridging dynamic graph attention network module to extract topological features of a brain region; processing the differential entropy sequence by using a multi-scale space-time mixed attention module to obtain multi-scale space-time features; carrying out residual mutual cross attention fusion on the topological features of the brain region and the multi-scale spatial-temporal features to obtain fusion features; and performing classification based on the fusion features, and determining an emotion recognition result corresponding to the electroencephalogram signal. According to the method, the accuracy and robustness of emotion recognition are improved by utilizing the spatial topology characteristics and the multi-topology time dynamic characteristics of the electroencephalogram signals, and the defects of modeling spatial dependence and time dynamic are overcome.

Provided are systems and methods for enhancing the output accuracy and consistency of sequence processing models, particularly when generating text that must conform to a specific formal language. Systems and methods can integrate finite-state or push-down automata with the decoding process to ensure that the generated sequences adhere to the desired syntax or grammar rules. A system can generate or otherwise leverage a finite-state or push-down automaton that encodes constraints associated with a particular formal language or syntax. At each decoding iteration of a sequence processing model, the decoding system can use the finite-state or push-down automaton to evaluate the token validity for each of a number of possible output tokens. The output of the sequence processing model can then be limited to or otherwise guided towards tokens that are indicated as valid by the finite-state or push-down automaton.

The invention relates to the technical field of system-on-chip service interrupt processing, in particular to a hardware Sequence-based SoC interrupt processing method, which comprises the following steps of: pre-programming an interrupt processing instruction sequence into an instruction random access memory (RAM), and configuring an instruction set, a sequence processor and an interrupt priority strategy of a task scheduler Sequence; receiving the interrupt signal by a task scheduler, and selecting a corresponding sequence processor according to an interrupt type; for the interrupt with low priority, the Sequencer directly executes the pre-stored instruction sequence to complete the interrupt response; and for high-priority interruption, the Sequencer caches the state of the register and sends a message packet to notify the CPU to perform cooperative processing. According to the method, interrupt autonomous response is achieved through hardware Sequence, zero CPU intervention is achieved in low-priority interrupt, a lightweight message mechanism is adopted in high-priority interrupt, and bandwidth occupation and power consumption are reduced while efficiency is improved.

Systems and methods for machine-learned generation of data insight summaries are provided. A computing system can obtain numerical time series data comprising a plurality of numerical values associated with a plurality of times. The computing system can identify, based on the numerical time series data, one or more first mathematical relationships in the numerical time series data. The computing system can generate, based at least in part on the mathematical relationships, a first input context comprising first natural language data indicative of the mathematical relationships. The computing system can provide the first input context to a first machine-learned sequence processing model. The first machine-learned sequence processing model can generate, based at least in part on the first input context, one or more outputs describing the one or more first mathematical relationships. The computing system can output the one or more outputs.

Provided are systems and methods for fine-tuning sequence processing models to human preferences. The approaches can account for tied preferences between pairs of sequences and, therefore, can be referred to as Tied Preference Optimization (TPO). Example sequence processing models include so-called large language models (LLMs), large multimodal models (LMMs), and other models that are configured to process inputs and / or generate outputs that are structured as a series of data elements such as tokens.

The invention belongs to the field of educational resource sharing, relates to a data analysis technology, is used for solving the problem that retrieval features of resources cannot be analyzed in the prior art, and particularly relates to an educational resource sharing system and method based on a cloud platform, and the system comprises a retrieval statistics module, a sharing processing module and a sharing evaluation module. The retrieval statistics module, the sharing processing module and the sharing evaluation module are in communication connection in sequence; the retrieval statistical module is used for performing statistical analysis on the educational resource retrieval data of the cloud platform: classifying the educational resources of the cloud platform to obtain a plurality of resource types, generating a statistical period, and marking a statistical data set of the resource types at the end moment of the statistical period; when the retrieval task is executed in the processing period, the processing output sequence is generated in combination with the keyword input by the user and the statistical data set of the corresponding processing type, the processing output sequence is composed of the conventional sub-sequence and the associated sub-sequence, and the retrieval efficiency of the user is improved.

The invention relates to the technical field of education intelligent analysis, and particularly discloses a learner cognition level fine-grained tracking method and system based on a state space model. The method aims at solving the problems that an existing knowledge tracking method is low in cognitive level modeling granularity, weak in long sequence processing capacity, poor in educational interpretation and the like. According to the method, a Bloom cognitive classification system and a state space modeling technology are combined, and fine-grained and multi-level dynamic modeling and future learning performance prediction of the knowledge mastering state of the learner are achieved. The core steps of the method comprise: constructing a semantic mapping relationship between knowledge points and cognitive hierarchies (S101); collecting and encoding multi-source learning behavior features (S102); learning a cognitive state evolution trajectory based on the state space model (S103); and outputting the cognitive hierarchy classification and the answer performance prediction (S104). According to the method, by fusing the multi-dimensional learning behavior data and the state space modeling capability, the accuracy and personalized analysis depth of cognitive tracking are improved, and technical support is provided for precise teaching and intelligent decision making.

Aspects of the disclosed technology include machine-learning systems and methods for generating user interface elements that allow user control over generative content creation by machine-learned generative models. A generative user interface (UI) system is configured to generate, as output of one or more machine-learned sequence processing models, computer-executable functional code to process a user query in association with a content item. The system is configured to generate computer-executable interface code for a user interface that includes a user interface element associated with at least one parameter of the computer-executable functional code for modifying the content item. The system is configured to determine data for the at least one parameter of the computer-executable functional code based at least in part on a user input to the user interface element and generate a modified content item using the computer-executable functional code and the data for the at least one parameter.

Aspects of the disclosed technology include computer-implemented systems and methods for conversational recommendation systems, such as conversational chatbots that are configured to process user queries and generate responses. A recommendation system includes a conversational user interface configured to receive a user query and provide a recommendation response and a machine-learned sequence processing model that has been trained on training data including a plurality of triplets. Each triplet includes an example query, an example model reasoning plan associated with the example query, and an example response associated with the example query and the example model reasoning plan. The sequence processing model can be trained to provide conversational-based recommendations using a multi-stage recommendation process that includes a planning stage, a conversation stage, and a retrieval stage.

The invention belongs to the technical field of education data mining and personalized learning, discloses a learner knowledge cognition level diagnosis method and system based on cross-scale learning performance dynamic modeling, and has higher accuracy in the aspects of learner cognition state prediction and knowledge point difficulty assessment. Through a selective state space modeling mechanism and cross-scale historical learning income feature engineering, cognitive change tracks of students in various learning scenes can be accurately captured; and the robustness, convergence efficiency and long sequence processing capability of the model in learner performance prediction are improved. The method can be widely applied to a personalized education platform, a self-adaptive learning system and an intelligent teaching auxiliary tool, provides accurate student learning state analysis for teachers, optimizes learning path design, and improves the teaching effect.

Provided are computing systems, methods, and platforms that rank text with pairwise ranking prompting using a generative sequence processing model. A prompt comprising a query and sets of text associated with candidate results can be generated. The generative sequence processing model can be prompted with the prompt and perform pairwise comparisons between the sets of text in the prompt based on the query in the prompt. An output can be generated that ranks the sets of text in response to the query.

The invention relates to the technical field of signal reconstruction, in particular to a complex electromagnetic environment-oriented millimeter wave signal blind source separation and reconstruction system, which comprises a frequency spectrum trend division module, a path fading construction module, an initial cluster label generation module, a multi-solution path screening module and a fusion reconstruction execution module. According to the method, power spectral density sequence processing is carried out on millimeter wave frequency domain data, a multi-dimensional feature group is formed according to a path loss factor, an angle of arrival and the like, and an initial feature cluster is screened through an Euclidean distance, so that the accuracy of signal source classification is effectively improved; after the frequency domain response and the phase contour are continuously subjected to point comparison, path screening is completed by integrating a mean square error residual value, the path misjudgment probability is reduced, time domain resampling and phase frequency offset standardization are executed in fusion reconstruction, the consistency and fidelity of signal reconstruction are enhanced, and the reconstruction precision is improved. The whole process improves the separation accuracy and reconstruction precision of mixed signals in a complex electromagnetic environment.

The invention belongs to the technical field of deep learning and brain science, and particularly relates to an ADHD graph convolution model construction method based on fMRI spatial-temporal characteristics, and the method comprises the following steps: processing resting state functional magnetic resonance data; fMRI sequence processing: randomly cutting fMRI sequences from different sites to obtain fMRI sequences with consistent sequence lengths, and obtaining a functional connection matrix according to Pearson's correlation; model input data: the model input data is composed of an fMRI sequence, ADHD phenotype information and a functional connection matrix, and the input data is divided to obtain a training set, a verification set and a test set; constructing a model; and model training: model evaluation. According to the method, the gated feature fusion module is proposed to fuse the captured fMRI features to obtain the spatial-temporal features, the phenotypic information of the subject is considered, the ADHD graph convolution model based on the fMRI spatial-temporal features is constructed, and the classification accuracy of ADHD diseases is improved.