Abstract:
Hierarchical and lightweight imitation learning (IL) for power management of embedded systems-on-chip (SoCs), also referred to herein as HiLITE, is provided. Modern SoCs use dynamic power management (DPM) techniques to improve energy efficiency. However, existing techniques are unable to efficiently adapt the runtime decisions considering multiple objectives (e.g., energy and real-time requirements) simultaneously on heterogeneous platforms. To address this need, embodiments described herein propose HiLITE, a hierarchical IL framework that maximizes energy efficiency while satisfying soft real-time constraints on embedded SoCs. This approach first trains DPM policies using IL; then, it applies a regression policy at runtime to minimize deadline misses. HiLITE improves the energy-delay product by 40% on average, and reduces deadline misses by up to 76%, compared to state-of-the-art approaches. In addition, the trained policies not only achieve high accuracy, but also have negligible prediction time overhead and small memory footprint.
Abstract:
Runtime task scheduling using imitation learning (IL) for heterogenous many-core systems is provided. Domain-specific systems-on-chip (DSSoCs) are recognized as a key approach to narrow down the performance and energy-efficiency gap between custom hardware accelerators and programmable processors. Reaching the full potential of these architectures depends critically on optimally scheduling the applications to available resources at runtime. Existing optimization-based techniques cannot achieve this objective at runtime due to the combinatorial nature of the task scheduling problem. In an exemplary aspect described herein, scheduling is posed as a classification problem, and embodiments propose a hierarchical IL-based scheduler that learns from an Oracle to maximize the performance of multiple domain-specific applications. Extensive evaluations show that the proposed IL-based scheduler approximates an offline Oracle policy with more than 99% accuracy for performance- and energy-based optimization objectives. Furthermore, it achieves almost identical performance to the Oracle with a low runtime overhead and high adaptivity.
Abstract:
A user-space emulation framework for heterogeneous system-on-chip (SoC) design is provided. Embodiments described herein propose a portable, Linux-based emulation framework to provide an ecosystem for hardware-software co-design of heterogeneous SoCs (e.g., domain-specific SoCs (DSSoCs)) and enable their rapid evaluation during the pre-silicon design phase. This framework holistically targets three key challenges of heterogeneous SoC design: accelerator integration, resource management, and application development. These challenges are addressed via a flexible and lightweight user-space runtime environment that enables easy integration of new accelerators, scheduling heuristics, and user applications, and the utility of each is illustrated through various case studies. A prototype compilation toolchain is introduced that enables automatic mapping of unlabeled C code to heterogeneous SoC platforms. Taken together, this environment offers a unique ecosystem to rapidly perform functional verification and obtain performance and utilization estimates that help accelerate convergence towards a final heterogeneous SoC design.
Abstract:
A photovoltaic device configured to substantially avoid radiative recombination of photo-generated carriers, reduce loss of energy of the photo-generated carriers through the phonon emission, extract photo-generated carriers substantially exclusively from the multi-frequency satellite valley(s) of the bandstructure of the used semiconductor material as opposed to the single predetermined extremum of the bandstructure. Methodologies of fabrication and operation of such a device.
Abstract:
Implementations of diagnostic systems for detecting a bacterial infection in a subject may include: a device for receiving a sample of a bodily fluid from a subject, a chemical composition configured to react with one or more antibodies in the bodily fluid, and an indicator configured to indicate or respond to a product of the reaction of the chemical composition with one or more antibodies in a bodily fluid received into the device. The antibodies may be produced in response to one or more conserved antigens from one or more bacteria identified as potentially associated with at least one disease associated with the subject. The device may be coated with the chemical composition. The indicator may be configured to communicate to a user of the system a presence or an absence of the one or more antibodies in the bodily fluid.
Abstract:
The invention is directed to methods for selecting a treatment option for an activated B cell-like diffuse large B cell lymphoma (ABC DLBCL) subject, a germinal center B cell-like diffuse large B cell lymphoma (GCB DLBCL) subject, a primary mediastinal B cell lymphoma (PMBL) subject, a Burkitt lymphoma (BL) subject, or a mantle cell lymphoma (MCL) subject by analyzing digital gene expression data obtained from the subject, e.g., from a biopsy sample.
Abstract:
Taught herein is a new contact-based optical imaging technology, en-face differential optical topography ( en-face DOT), which performs real-time visualization of subsurface tissue heterogeneity within a depth up to 3mm and over a 9.5mm diameter FOV with a modest mm-level lateral resolution. An embodiment of the probe fits in a 12mm port and houses at its maximum 128 copper-coated 750μηι fibers that form radially alternating illumination (70 fibers) and detection (58 fibers) channels. By simultaneously illuminating the 70 source channels of the laparoscopic probe that is in contact with a scattering medium and concurrently measuring the light diffusely propagated to the 58 detector channels, the presence of near-surface optical heterogeneities can be resolved in an en-face 9.5mm field-of-view in real-time. Visualization of subsurface margin of strong attenuation contrast at a depth up to 3 mm is demonstrated at one wavelength at a frame rate of 1.3 Hz.
Abstract:
Identification and evaluation of a set of first-in-class potent inhibitors targeting a new cancer target, Grb2-associated binder~1 (GAB1 ), which integrates signals from different signaling pathways and is frequently over-expressed in cancer ceils. Intensive computational modeling is utilized to understand the structure of the GAB1 pleckstrin homology (PH) domain and screened five million compounds. Upon biological evaluation, several inhibitors were found that induced large conformational changes of the target structure exhibited strong selective binding to GAB1 PH domain. Particularly, these inhibitors demonstrated potent and tumor-specific cytotoxicity in breast cancer cells. This targeting GAB1 signaling may be used for cancer therapy, especially for triple negative breast cancer patients.
Abstract:
Methods and compositions for the prognosis and classification of cancer, especially breast cancer, are provided. For example, in certain aspects methods for cancer prognosis using copy number analysis of selected biomarkers are described. In further aspects, copy number information may be used to predict metastasis risk and help treatment decision-making.