Intel Corporation

Anirud Thyagharajan

Om Ji Omer

Sreenivas Subramoney

Gurpreet Singh Kalsi

Kamlesh R Pillai

Dipan Kumar Mandal

Prashant Laddha

Gopi Neela

Avishaii Abuhatzera

Benjamin Ummenhofer

Srivatsa Rangachar Srinivasa

Saurabh Jain

Srivatsava Jandhyala

Krishnan Ananthanarayanan

Akshay Krishna Ramanathan

Meenal Kudalkar

A memory architecture includes processing circuits co-located with memory subarrays for performing computations within the memory architecture. The memory architecture includes a plurality of decoders in hierarchical levels that include a multicast capability for distributing data or compute operations to individual subarrays. The multicast may be configurable with respect to individual fan-outs at each hierarchical level. A computation workflow may be organized into a compute supertile representing one or more 'supertiles' of input data to be processed in the compute supertile. The individual data tiles of the input data supertile may be used by multiple compute tiles executed by the processing circuits of the subarrays, and the data tiles multicast to the respective processing circuits for efficient data loading and parallel computation.

Hardware-software co-designed multi-cast for in-memory computing architectures

An embodiment of an apparatus comprises a hardware accelerator to perform a three-dimensional (3D) point cloud data access operation, and circuitry coupled to the hardware accelerator to control the hardware accelerator to perform the 3D point cloud data access operation in response to a request. Other embodiments are disclosed and claimed.

Point cloud adjacency-map and hash-map accelerator

Exemplary embodiments maintain spatial locality of the data being processed by a sparse CNN. The spatial locality is maintained by reordering the data to preserve spatial locality. The reordering may be performed on data elements and on data for groups of co-located data elements referred to herein as 'chunks'. Thus, the data may be reordered into chunks, where each chunk contains data for spatially co-located data elements, and in addition, chunks may be organized so that spatially located chunks are together. The use of chunks helps to reduce the need to re-fetch data during processing. Chunk sizes may be chosen based on the memory constraints of the processing logic (e.g., cache sizes).

Methods, apparatus, and articles of manufacture to reorder N-dimensional sparse data into groups of data elements that can be collocated in a memory

Methods, apparatus, systems, and articles of manufacture are disclosed to improve in-memory multiply and accumulate operations. An example apparatus includes a first multiplexer in a subarray of memory, the first multiplexer to receive first values representative of a column of a lookup table (LUT) including entries to represent products of four-bit numbers and return second values from an intersection of a row and the column of the LUT based on a first element of a first operand; shift and adder logic in the subarray, the shift and adder logic to shift the second values based on at least one of the first element of the first operand or a first element of a second operand; and accumulation storage in the subarray, the accumulation storage to store at least the shifted second values.

Methods, apparatus, and articles of manufacture to improve in-memory multiply and accumulate operations

A computer model is trained to classify regions of a space (e.g., a pixel of an image or a voxel of a point cloud) according to a multi-label classification. To improve the model's accuracy, the model's self-confidence is determined with respect to its own predictions of regions in a training space. The self-confidence is determined based on the class predictions, such as a difference between the highest-predicted class and a second-highest-predicted class. When these are similar, it may reflect areas for potential improvement by focusing training on these low-confidence areas. Additional training may be performed by including modified training data in subsequent training iterations that focuses on low-confidence areas. As another example, additional training may be performed using the self-confidence to modify a classification loss used to refine parameters of the model.

Applying self-confidence in multi-label classification to model training

Techniques related to indirect sparse simultaneous localization and mapping (SLAM) are discussed. Such techniques include adaptively positioning a virtual camera relative to an estimated position of a physical camera within an environment to be mapped, projecting a depth error to an image plane corresponding to the adaptive camera position, and using the projected depth error to update a mapping of the environment.

Adaptive virtual camera for indirect-sparse simultaneous localization and mapping systems

An embodiment of an image processor device includes technology to fetch a feature point data set from outside a local memory, locally store three or more fetched feature point data sets in the local memory, compute orientation information for each fetched feature point data set, compute first descriptor information based on the computed orientation information and a first locally stored feature point data set in parallel with a fetch and local store of a second feature point data set in the local memory, and compute second descriptor information based on the computed orientation information and the second locally stored feature point data set in parallel with the compute of the first descriptor information. Other embodiments are disclosed and claimed.

Architecture to generate binary descriptor for image feature point

An apparatus, method, and computer readable medium to accommodate depth noise in SLAM (Simultaneous Localization & Mapping). The method includes receiving correspondences and clusters for pose estimation. After the correspondences and clusters are received, a dynamic centroid is determined using 3-D features of a landmark. The 3-D features of the landmark comprise a cluster. Next, distance-error metrics are determined using the dynamic centroid, a map point, and the cluster. The distance-error metrics are compared with thresholds to remove depth noise affected 3-D features and landmarks when the distance-error metrics are larger than the thresholds. The remaining 3-D features of the cluster are sent to a pose estimation framework.

Growing community of inventors

Bengaluru, India

Anirud Thyagharajan