Abstract: Machine learning-based techniques for summarizing collections of data such as image and video data leveraging side information obtained from related (e.g., video) data are provided. In one aspect, a method for video summarization includes: obtaining related videos having content related to a target video; and creating a summary of the target video using information provided by the target video and side information provided by the related videos to select portions of the target video to include in the summary. The side information can include video data, still image data, text, comments, natural language descriptions, and combinations thereof.

Abstract: A processor training a reinforcement learning model can include receiving a first dataset representing an observable state in reinforcement learning to train a machine to perform an action. The processor receives a second dataset. Using the second dataset, the processor trains a machine learning classifier to make a prediction about an entity related to the action. The processor extracts an embedding from the trained machine learning classifier, and augments the observable state with the embedding to create an augmented state. Based on the augmented state, the processor trains a reinforcement learning model to learn a policy for performing the action, the policy including a mapping from state space to action space.

Abstract: In an embodiment, a method for generating contrastive information for a classifier prediction comprises receiving image data representative of an input image, using a deep learning classifier model to predict a first classification for the input image, evaluating the input image using a plurality of classifier functions corresponding to respective high-level features to identify one or more of the high-level features absent from the input image, and identifying, from among the high-level features absent from the input image, a pertinent-negative feature that, if added to the input image, will result in the deep learning classifier model predicting a second classification for the modified input image, the second classification being different from the first classification. In an embodiment, the method includes creating a pertinent-positive image that is a modified version of the input image that has the first classification and fewer than all superpixels of the input image.

Abstract: Transfer learning in machine learning can include receiving a machine learning model. Target domain training data for reprogramming the machine learning model using transfer learning can be received. The target domain training data can be transformed by performing a transformation function on the target domain training data. Output labels of the machine learning model can be mapped to target labels associated with the target domain training data. The transformation function can be trained by optimizing a parameter of the transformation function. The machine learning model can be reprogrammed based on input data transformed by the transformation function and a mapping of the output labels to target labels.

Abstract: A query-based generic end-to-end molecular optimization (“QMO”) system framework, method and computer program product for optimizing molecules, such as for accelerating drug discovery. The QMO framework decouples representation learning and guided search and applies to any plug-in encoder-decoder with continuous latent representations. QMO framework directly incorporates evaluations based on chemical modeling, analysis packages, and pre-trained machine-learned prediction models for efficient molecule optimization using a query-based guided search method based on zeroth order optimization. The QMO features efficient guided search with molecular property evaluations and constraints obtained using the predictive models and chemical modeling and analysis packages.

Abstract: One or more computer processors generate a plurality of adversarial perturbations associated with a model, wherein the plurality of adversarial perturbations comprises a universal perturbation and one or more per-sample perturbations. The one or more computer processors identify a plurality of neuron activations associated with the model and the plurality of generated adversarial perturbations. The one or more computer processors maximize the identified plurality of neuron activations. The one or more computer processors determine the model is a Trojan model by leveraging one or more similarities associated with the maximized neuron activations and the generated adversarial perturbations.

Abstract: A method and system of analyzing a symbolic sequence is provided. Metadata of a symbolic sequence is received from a computing device of an owner. A set of R random sequences are generated based on the received metadata and sent to the computing device of the owner of the symbolic sequence for computation of a feature matrix based on the set of R random sequences and the symbolic sequence. The feature matrix is received from the computing device of the owner. Upon determining that an inner product of the feature matrix is below a threshold accuracy, the iterative process returns to generating R random sequences. Upon determining that the inner product of the feature matrix is at or above the threshold accuracy, the feature matrix is categorized based on machine learning. The categorized global feature matrix is sent to be displayed on a user interface of the computing device of the owner.

Abstract: With at least one hardware processor, obtain data specifying: two trained neural network models; and alignment data. With the at least one hardware processor, carry out neuron alignment on the two trained neural network models using the alignment data to obtain two aligned models. With the at least one hardware processor, train a minimal loss curve between the two aligned models. With the at least one hardware processor, select a new model along the minimal loss curve that maximizes accuracy on adversarially perturbed data.

Abstract: In an embodiment, a method for generating contrastive information for a classifier prediction comprises receiving image data representative of an input image, using a deep learning classifier model to predict a first classification for the input image, evaluating the input image using a plurality of classifier functions corresponding to respective high-level features to identify one or more of the high-level features absent from the input image, and identifying, from among the high-level features absent from the input image, a pertinent-negative feature that, if added to the input image, will result in the deep learning classifier model predicting a second classification for the modified input image, the second classification being different from the first classification. In an embodiment, the method includes creating a pertinent-positive image that is a modified version of the input image that has the first classification and fewer than all superpixels of the input image.

Abstract: Techniques for filtering artificial intelligence (AI)-designed molecules for laboratory testing provided. According to an embodiment, computer implemented method can comprise selecting, by a system operatively coupled to a processor, a first subset of AI-designed molecules from a set of AI-designed molecules as candidate pharmaceutical agents based on classification of the AI-designed molecules using one or more classifiers. The method further comprises selecting, by the system, a second subset of the candidate pharmaceutical agents for wet laboratory testing based on evaluation of molecular interactions between the candidate pharmaceutical agents and one or more biological targets using one or more computer simulations.

Abstract: A method of utilizing a computing device to optimize weights within a neural network to avoid adversarial attacks includes receiving, by a computing device, a neural network for optimization. The method further includes determining, by the computing device, on a region by region basis one or more robustness bounds for weights within the neural network. The robustness bounds indicating values beyond which the neural network generates an erroneous output upon performing an adversarial attack on the neural network. The computing device further averages all robustness bounds on the region by region basis. The computing device additionally optimizes weights for adversarial proofing the neural network based at least in part on the averaged robustness bounds.

Abstract: Techniques for training robust machine learning models for adversarial input data. Training data for a machine learning (ML) model is received. The training data includes a plurality of labels for data elements. First modified training data is generated by modifying one or more of the plurality of labels in the training data using parameterized label smoothing with a first optimization parameter. The ML model is trained using the first modified training data.

Abstract: An optical fingerprint sensing module attached to a base is provided. The base includes a first surface, a second surface and an opening extending through the first surface and the second surface. The optical fingerprint sensing module includes a fixing frame and a sensor integrated circuit (IC). The fixing frame is disposed in the opening of the base. The sensor IC is disposed in a receiving groove of the fixing frame and includes a plurality of photo sensors. The photo sensors receive light reflected from a user's finger through the opening of the base.

Abstract: Embodiments relate to a system, program product, and method for inducing creativity in an artificial neural network (ANN) having an encoder and decoder. Neurons are automatically selected and manipulated from one or more layers of the encoder. An encoded vector is sampled for an encoded image. Decoder neurons and a corresponding activation pattern are evaluated with respect to the encoded image. The decoder neurons that correspond to the activation pattern are selected, and an activation setting of the selected decoder neurons is changed. One or more novel data instances are automatically generated from an original latent space of the selectively changed decoder neurons.

Abstract: A computer implemented method and system of presenting content on a display of a computing device is provided. Historical data including data of a plurality of customers and data of a plurality of products is received. A hybrid graph is created. The hybrid graph includes one or more customer nodes and product nodes. Between each two customers, a customer weight factor is applied. Between each two products, a product weight factor is applied. One or more products related to a seed product are identified. For each identified related product, a return affinity score towards the requestor customer is determined. A representation of the related products is sent to be displayed on the computing device of the requestor customer, based on the affinity score.

Abstract: The present invention relates to an electronic device with a lens supporting element. The electronic device includes a display screen, an intermediate frame, a lens supporting element, an interface structure and a base structure. The intermediate frame has a frame hollow region. A fingerprint sensing unit is installed on the lens supporting element. The fingerprint sensing unit is located near the frame hollow region through the lens supporting element. The interface structure is arranged between the intermediate frame and the lens supporting element, and/or arranged between the lens supporting element and the base frame. Consequently, the lens supporting element is positioned in or clamped between the intermediate frame and the base structure through the interface structure.

Abstract: Embodiments relate to a system, program product, and method to support a convolutional neural network (CNN). A class-specific discriminative image region is localized to interpret a prediction of a CNN and to apply a class activation map (CAM) function to received input data. First and second attacks are generated on the CNN with respect to the received input data. The first attack generates first perturbed data and a corresponding first CAM, and the second attack generates second perturbed data and a corresponding second CAM. An interpretability discrepancy is measured to quantify one or more differences between the first CAM and the second CAM. The measured interpretability discrepancy is applied to the CNN. The application is a response to an inconsistency between the first CAM and the second CAM and functions to strengthen the CNN against an adversarial attack.

Abstract: An electronic device is provided, including a housing, a front cover, a display panel module, a conductive structure, a circuit board, and an IC element, wherein the front cover is disposed on a front side of the electronic device and coupled to the housing. The conductive structure is disposed on the front cover and located outside an active display region of the display panel module. The IC element is disposed on the circuit board and electrically coupled to the conductive structure. When a human body is approaching to the front cover, the conductive structure generates an electrical signal to the IC element.

Abstract: Methods, systems, and circuits for training a neural network include applying noise to a set of training data across wordlines using a respective noise switch on each wordline. A neural network is trained using the noise-applied training data to generate a classifier that is robust against adversarial training.

Abstract: A method, computer system, and a computer program product for designing one or more folded structural proteins from at least one raw amino acid sequence is provided. The present invention may include computing one or more character embeddings based on the at least one raw amino acid sequence by utilizing a multi-scale neighborhood-based neural network (MNNN) model. The present invention may then include refining the computed one or more character embeddings with at least one set of sequence neighborhood information. The present invention may further include predicting one or more dihedral angles based on the refined one or more character embeddings.