Electrical and Computer Engineering

Speech signals are valuable biomarkers for assessing an individuals mental health, including identifying Major Depressive Disorder (MDD) automatically. A frequently used approach in this regard is to employ features related to speaker identity, such as speaker-embeddings. However, over-reliance on speaker identity features in mental health screening systems can compromise patient privacy. Moreover, some aspects of speaker identity may not be relevant for depression detection and could serve as a bias factor that hampers system performance. To overcome these limitations, we propose disentangling speaker-identity information from depression-related information. Specifically, we present four distinct disentanglement methods to achieve this - adversarial speaker identification (SID)-loss maximization (ADV), SID-loss equalization with variance (LEV), SID-loss equalization using Cross-Entropy (LECE) and SID-loss equalization using KL divergence (LEKLD). Our experiments, which incorporated diverse input features and model architectures, have yielded improved F1 scores for MDD detection and voice-privacy attributes, as quantified by Gain in Voice Distinctiveness GV D and De-Identification Scores (DeID). On the DAIC-WOZ dataset (English), LECE using ComparE16 features results in the best F1-Scores of 80% which represents the audio-only SOTA depression detection F1-Score along with a GV D of -1.1 dB and a DeID of 85%. On the EATD dataset (Mandarin), ADV using raw-audio signal achieves an F1-Score of 72.38% surpassing multi-modal SOTA along with a GV D of -0.89 dB dB and a DeID of 51.21%. By reducing the dependence on speaker-identity-related features, our method offers a promising direction for speech-based depression detection that preserves patient privacy.

Objective. This study aims to develop and validate an end-to-end software platform, PyHFO, that streamlines the application of deep learning (DL) methodologies in detecting neurophysiological biomarkers for epileptogenic zones from EEG recordings.Approach. We introduced PyHFO, which enables time-efficient high-frequency oscillation (HFO) detection algorithms like short-term energy and Montreal Neurological Institute and Hospital detectors. It incorporates DL models for artifact and HFO with spike classification, designed to operate efficiently on standard computer hardware.Main results. The validation of PyHFO was conducted on three separate datasets: the first comprised solely of grid/strip electrodes, the second a combination of grid/strip and depth electrodes, and the third derived from rodent studies, which sampled the neocortex and hippocampus using depth electrodes. PyHFO demonstrated an ability to handle datasets efficiently, with optimization techniques enabling it to achieve speeds up to 50 times faster than traditional HFO detection applications. Users have the flexibility to employ our pre-trained DL model or use their EEG data for custom model training.Significance. PyHFO successfully bridges the computational challenge faced in applying DL techniques to EEG data analysis in epilepsy studies, presenting a feasible solution for both clinical and research settings. By offering a user-friendly and computationally efficient platform, PyHFO paves the way for broader adoption of advanced EEG data analysis tools in clinical practice and fosters potential for large-scale research collaborations.

The effects of optical feedback on a terahertz (THz) quantum-cascade metasurface vertical-external-cavity surface-emitting laser (QC-VECSEL) are investigated via self-mixing. A single-mode 2.80 THz QC-VECSEL operating in continuous-wave is subjected to various optical feedback conditions (i.e., feedback strength, round-trip time, and angular misalignment) while variations in its terminal voltage associated with self-mixing are monitored. Due to its large radiating aperture and near-Gaussian beam shape, we find that the QC-VECSEL is strongly susceptible to optical feedback, which is robust against misalignment of external optics. This, in addition to the use of a high-reflectance flat output coupler, results in high feedback levels associated with multiple round-trips within the external cavity-a phenomenon not typically observed for ridge-waveguide QC-lasers. Thus, a new theoretical model is established to describe self-mixing in the QC-VECSEL. The stability of the device under variable optical feedback conditions is also studied. Any mechanical instabilities of the external cavity (such as vibrations of the output coupler), are enhanced due to feedback and result in low-frequency oscillations of the terminal voltage. The work reveals how the self-mixing response differs for the QC-VECSEL architecture, informs other systems in which optical feedback is unavoidable, and paves the way for QC-VECSEL self-mixing applications.

Machine learning at the extreme edge has enabled a plethora of intelligent, time-critical, and remote applications. However, deploying interpretable artificial intelligence systems that can perform high-level symbolic reasoning and satisfy the underlying system rules and physics within the tight platform resource constraints is challenging. In this paper, we introduce TinyNS, the first platform-aware neurosymbolic architecture search framework for joint optimization of symbolic and neural operators. TinyNS provides recipes and parsers to automatically write microcontroller code for five types of neurosymbolic models, combining the context awareness and integrity of symbolic techniques with the robustness and performance of machine learning models. TinyNS uses a fast, gradient-free, black-box Bayesian optimizer over discontinuous, conditional, numeric, and categorical search spaces to find the best synergy of symbolic code and neural networks within the hardware resource budget. To guarantee deployability, TinyNS talks to the target hardware during the optimization process. We showcase the utility of TinyNS by deploying microcontroller-class neurosymbolic models through several case studies. In all use cases, TinyNS outperforms purely neural or purely symbolic approaches while guaranteeing execution on real hardware.

The InGaAs lattice-matched to InP has been widely deployed as the absorption material in short-wavelength infrared photodetection applications such as imaging and optical communications. Here, a series of digital alloy (DA)-grown InAs/GaAs short-period superlattices were investigated to extend the absorption spectral range. The scanning transmission electron microscopy, high-resolution X-ray diffraction, and atomic force microscopy measurements exhibit good material quality, while the photoluminescence (PL) spectra demonstrate a wide band gap tunability for the InGaAs obtained via the DA growth technique. The photoluminescence peak can be effectively shifted from 1690 nm (0.734 eV) for conventional random alloy (RA) InGaAs to 1950 nm (0.636 eV) for 8 monolayer (ML) DA InGaAs at room temperature. The complete set of optical constants of DA InGaAs has been extracted via the ellipsometry technique, showing the absorption coefficients of 398, 831, and 1230 cm-1 at 2 μm for 6, 8, and 10 ML DA InGaAs, respectively. As the period thickness increases for DA InGaAs, a red shift at the absorption edge can be observed. Furthermore, the simulated band structures of DA InGaAs via an environment-dependent tight binding model agree well with the measured photoluminescence peaks, which is advantageous for a physical understanding of band structure engineering via the DA growth technique. These investigations and results pave the way for the future utilization of the DA-grown InAs/GaAs short-period superlattices as a promising absorption material choice to extend the photodetector response beyond the cutoff wavelength of random alloy InGaAs.

Design strategies for improving terahertz (THz) quantum cascade lasers (QCLs) in the 5-6 THz range are investigated numerically and experimentally, with the goal of overcoming the degradation in performance that occurs as the laser frequency approaches the Reststrahlen band. Two designs aimed at 5.4 THz were selected: one optimized for lower power dissipation and one optimized for better temperature performance. The active regions exhibited broadband gain, with the strongest modes lasing in the 5.3-5.6 THz range, but with other various modes observed ranging from 4.76 to 6.03 THz. Pulsed and continuous-wave (cw) operation is observed up to temperatures of 117 K and 68 K, respectively. In cw mode, the ridge laser has modes up to 5.71 THz - the highest reported frequency for a THz QCL in cw mode. The waveguide loss associated with the doped contact layers and metallization is identified as a critical limitation to performance above 5 THz.

Image denoising, one of the essential inverse problems, targets to remove noise/artifacts from input images. In general, digital image denoising algorithms, executed on computers, present latency due to several iterations implemented in, e.g., graphics processing units (GPUs). While deep learning-enabled methods can operate non-iteratively, they also introduce latency and impose a significant computational burden, leading to increased power consumption. Here, we introduce an analog diffractive image denoiser to all-optically and non-iteratively clean various forms of noise and artifacts from input images - implemented at the speed of light propagation within a thin diffractive visual processor that axially spans <250 × λ, where λ is the wavelength of light. This all-optical image denoiser comprises passive transmissive layers optimized using deep learning to physically scatter the optical modes that represent various noise features, causing them to miss the output image Field-of-View (FoV) while retaining the object features of interest. Our results show that these diffractive denoisers can efficiently remove salt and pepper noise and image rendering-related spatial artifacts from input phase or intensity images while achieving an output power efficiency of ~30-40%. We experimentally demonstrated the effectiveness of this analog denoiser architecture using a 3D-printed diffractive visual processor operating at the terahertz spectrum. Owing to their speed, power-efficiency, and minimal computational overhead, all-optical diffractive denoisers can be transformative for various image display and projection systems, including, e.g., holographic displays.

Bremsstrahlung-the spontaneous emission of broadband radiation from free electrons that are deflected by atomic nuclei-contributes to the majority of X-rays emitted from X-ray tubes and used in applications ranging from medical imaging to semiconductor chip inspection. Here, we show that the bremsstrahlung intensity can be enhanced significantly-by more than three orders of magnitude-through shaping the electron wavefunction to periodically overlap with atoms in crystalline materials. Furthermore, we show how to shape the bremsstrahlung X-ray emission pattern into arbitrary angular emission profiles for purposes such as unidirectionality and multi-directionality. Importantly, we find that these enhancements and shaped emission profiles cannot be attributed solely to the spatial overlap between the electron probability distribution and the atomic centers, as predicted by the paraxial and non-recoil theory for free electron light emission. Our work highlights an unprecedented regime of free electron light emission where electron waveshaping provides multi-dimensional control over practical radiation processes like bremsstrahlung. Our results pave the way towards greater versatility in table-top X-ray sources and improved fundamental understanding of quantum electron-light interactions.

Mitosis is a critical criterion for meningioma grading. However, pathologists assessment of mitoses is subject to significant inter-observer variation due to challenges in locating mitosis hotspots and accurately detecting mitotic figures. To address this issue, we leverage digital pathology and propose a computational strategy to enhance pathologists mitosis assessment. The strategy has two components: (1) A depth-first search algorithm that quantifies the mathematically maximum mitotic count in 10 consecutive high-power fields, which can enhance the preciseness, especially in cases with borderline mitotic count. (2) Implementing a collaborative sphere to group a set of pathologists to detect mitoses under each high-power field, which can mitigate subjective random errors in mitosis detection originating from individual detection errors. By depth-first search algorithm (1) , we analyzed 19 meningioma slides and discovered that the proposed algorithm upgraded two borderline cases verified at consensus conferences. This improvement is attributed to the algorithms ability to quantify the mitotic count more comprehensively compared to other conventional methods of counting mitoses. In implementing a collaborative sphere (2) , we evaluated the correctness of mitosis detection from grouped pathologists and/or pathology residents, where each member of the group annotated a set of 48 high-power field images for mitotic figures independently. We report that groups with sizes of three can achieve an average precision of 0.897 and sensitivity of 0.699 in mitosis detection, which is higher than an average pathologist in this study (precision: 0.750, sensitivity: 0.667). The proposed computational strategy can be integrated with artificial intelligence workflow, which envisions the future of achieving a rapid and robust mitosis assessment by interactive assisting algorithms that can ultimately benefit patient management.

Inverse design is a powerful tool in wave physics for compact, high-performance devices. To date, applications in photonics have mostly been limited to linear systems and it has rarely been investigated or demonstrated in the nonlinear regime. In addition, the “black box” nature of inverse design techniques has hindered the understanding of optimized inverse-designed structures. We propose an inverse design method with interpretable results to enhance the efficiency of on-chip photon generation rate through nonlinear processes by controlling the effective phase-matching conditions. We fabricate and characterize a compact, inverse-designed device using a silicon-on-insulator platform that allows a spontaneous four-wave mixing process to generate photon pairs at a rate of 1.1 MHz with a coincidence to accidental ratio of 162. Our design method accounts for fabrication constraints and can be used for scalable quantum light sources in large-scale communication and computing applications.