Publications without Fulltext
Permanent URI for this collectionhttps://hdl.handle.net/20.500.14288/3
Browse
1014 results
Filters
Advanced Search
Filter by
Settings
Search Results
Publication Metadata only Implementing the analogous neural network using chaotic strange attractors(Springer Nature, 2024) Department of Electrical and Electronics Engineering; Teğin, Uğur; Department of Electrical and Electronics Engineering; College of EngineeringMachine learning studies need colossal power to process massive datasets and train neural networks to reach high accuracies, which have become gradually unsustainable. Limited by the von Neumann bottleneck, current computing architectures and methods fuel this high power consumption. Here, we present an analog computing method that harnesses chaotic nonlinear attractors to perform machine learning tasks with low power consumption. Inspired by neuromorphic computing, our model is a programmable, versatile, and generalized platform for machine learning tasks. Our mode provides exceptional performance in clustering by utilizing chaotic attractors’ nonlinear mapping and sensitivity to initial conditions. When deployed as a simple analog device, it only requires milliwatt-scale power levels while being on par with current machine learning techniques. We demonstrate low errors and high accuracies with our model for regression and classification-based learning tasks.Publication Metadata only Event-triggered reinforcement learning based joint resource allocation for ultra-reliable low-latency V2X communications(Institute of Electrical and Electronics Engineers Inc., 2024) Department of Electrical and Electronics Engineering; Ergen, Sinem Çöleri; Khan, Nasir; Department of Electrical and Electronics Engineering; College of Engineering; Graduate School of Sciences and EngineeringFuture 6G-enabled vehicular networks face the challenge of ensuring ultra-reliable low-latency communication (URLLC) for delivering safety-critical information in a timely manner. Existing resource allocation schemes for vehicle-toeverything (V2X) communication systems primarily rely on traditional optimization-based algorithms. However, these methods often fail to guarantee the strict reliability and latency requirements of URLLC applications in dynamic vehicular environments due to the high complexity and communication overhead of the solution methodologies. This paper proposes a novel deep reinforcement learning (DRL) based framework for the joint power and block length allocation to minimize the worst-case decoding-error probability in the finite block length (FBL) regime for a URLLC-based downlink V2X communication system. The problem is formulated as a non-convex mixed-integer nonlinear programming problem (MINLP). Initially, an algorithm grounded in optimization theory is developed based on deriving the joint convexity of the decoding error probability in the block length and transmit power variables within the region of interest. Subsequently, an efficient event-triggered DRL based algorithm is proposed to solve the joint optimization problem. Incorporating event-triggered learning into the DRL framework enables assessing whether to initiate the DRL process, thereby reducing the number of DRL process executions while maintaining reasonable reliability performance. The DRL framework consists of a twolayered structure. In the first layer, multiple deep Q-networks (DQNs) are established at the central trainer for block length optimization. The second layer involves an actor-critic network and utilizes the deep deterministic policy-gradient (DDPG)-based algorithm to optimize the power allocation. Simulation results demonstrate that the proposed event-triggered DRL scheme can achieve 95% of the performance of the joint optimization scheme while reducing the DRL executions by up to 24% for different network settings.Publication Metadata only Vilma: a zero-shot benchmark for linguıstic and temporal grounding in video-language models(International Conference on Learning Representations, ICLR, 2024) Pedrotti, Andrea; Dogan, Mustafa; Cafagna, Michele; Parcalabescu, Letitia; Calixto, Iacer; Frank, Anetteh; Gatt, Albert; Department of Electrical and Electronics Engineering; Kesen, İlker; Erdem, Aykut; Department of Electrical and Electronics Engineering; Koç Üniversitesi İş Bankası Yapay Zeka Uygulama ve Araştırma Merkezi (KUIS AI)/ Koç University İş Bank Artificial Intelligence Center (KUIS AI); Graduate School of Sciences and Engineering; College of EngineeringWith the ever-increasing popularity of pretrained Video-Language Models (VidLMs), there is a pressing need to develop robust evaluation methodologies that delve deeper into their visio-linguistic capabilities. To address this challenge, we present VILMA), a task-agnostic benchmark that places the assessment of fine-grained capabilities of these models on a firm footing. Task-based evaluations, while valuable, fail to capture the complexities and specific temporal aspects of moving images that VidLMs need to process. Through carefully curated counterfactuals, VILMA offers a controlled evaluation suite that sheds light on the true potential of these models, as well as their performance gaps compared to human-level understanding. VILMA also includes proficiency tests, which assess basic capabilities deemed essential to solving the main counterfactual tests. We show that current VidLMs' grounding abilities are no better than those of vision-language models which use static images. This is especially striking once the performance on proficiency tests is factored in. Our benchmark serves as a catalyst for future research on VidLMs, helping to highlight areas that still need to be explored.Publication Metadata only Microfluidic pulse shaping methods for molecular communications(Elsevier, 2023) Department of Electrical and Electronics Engineering; Kahvazi Zadeh, Maryam; Bolhassan, Iman Mokari; Kuşcu, Murat; Department of Electrical and Electronics Engineering; Graduate School of Sciences and Engineering; College of EngineeringMolecular Communication (MC) is a bio-inspired communication modality that utilizes chemical signals in the form of molecules to exchange information between spatially separated entities. Pulse shaping is an important process in all communication systems, as it modifies the waveform of transmitted signals to match the characteristics of the communication channel for reliable and high-speed information transfer. In MC systems, the unconventional architectures of components, such as transmitters and receivers, and the complex, nonlinear, and time-varying nature of MC channels make pulse shaping even more important. While several pulse shaping methods have been theoretically proposed for MC, their practicality and performance are still uncertain. Moreover, the majority of recently proposed experimental MC testbeds that rely on microfluidics technology lack the incorporation of programmable pulse shaping methods, which hinders the accurate evaluation of MC techniques in practical settings. To address the challenges associated with pulse shaping in microfluidic MC systems, we provide a comprehensive overview of practical microfluidic chemical waveform generation techniques that have been experimentally validated and whose architectures can inform the design of pulse shaping methods for microfluidic MC systems and testbeds. These techniques include those based on hydrodynamic and acoustofluidic force fields, as well as electrochemical reactions. We also discuss the fundamental working mechanisms and system architectures of these techniques, and compare their performances in terms of spatiotemporal resolution, selectivity, system complexity, and other performance metrics relevant to MC applications, as well as their feasibility for practical MC applications.Publication Metadata only Multi-scale deformable alignment and content-adaptive inference for flexible-rate bi-directional video compression(IEEE Computer Society, 2023) Department of Electrical and Electronics Engineering; Yılmaz, Mustafa Akın; Ulaş, Ökkeş Uğur; Tekalp, Ahmet Murat; Department of Electrical and Electronics Engineering; Graduate School of Sciences and Engineering; College of EngineeringThe lack of ability to adapt the motion compensation model to video content is an important limitation of current end-to-end learned video compression models. This paper advances the state-of-the-art by proposing an adaptive motion-compensation model for end-to-end rate-distortion optimized hierarchical bi-directional video compression. In particular, we propose two novelties: i) a multi-scale deformable alignment scheme at the feature level combined with multi-scale conditional coding, ii) motion-content adaptive inference. In addition, we employ a gain unit, which enables a single model to operate at multiple rate-distortion operating points. We also exploit the gain unit to control bit allocation among intra-coded vs. bi-directionally coded frames by fine tuning corresponding models for truly flexible-rate learned video coding. Experimental results demonstrate state-of-the-art rate-distortion performance exceeding those of all prior art in learned video coding1.Publication Metadata only Ris-aided angular-based hybrid beamforming design in mmwave massive mimo systems(IEEE, 2022) Koc, Asil; Tho Le-Ngoc; Department of Electrical and Electronics Engineering; Yıldırım, İbrahim; Başar, Ertuğrul; Department of Electrical and Electronics Engineering; Graduate School of Sciences and Engineering; College of EngineeringThis paper proposes a reconfigurable intelligent surface (RIS)-aided and angular-based hybrid beamforming (AB-HBF) technique for the millimeter wave (mmWave) massive multiple-input multiple-output (MIMO) systems. The proposed RIS-AB-HBF architecture consists of three stages: (i) RF beam-former, (ii) baseband (BB) precoder/combiner, and (iii) RIS phase shift design. First, in order to reduce the number of RF chains and the channel estimation overhead, RF beamformers are designed based on the 3D geometry-based mmWave channel model using slow time-varying angular parameters of the channel. Second, a BB precoder/combiner is designed by exploiting the reduced-size effective channel seen from the BB stages. Then, the phase shifts of the RIS are adjusted to maximize the achievable rate of the system via the nature-inspired particle swarm optimization (PSO) algorithm. Illustrative simulation results demonstrate that the use of RISs in the AB-HBF systems has the potential to provide more promising advantages in terms of reliability and flexibility in system design.Publication Metadata only IDE-integrated microneedle arrays as fully biodegradable platforms for wearable/implantable capacitive biosensing(Institute of Electrical and Electronics Engineers Inc., 2023) Department of Electrical and Electronics Engineering; Ürey, Hakan; Mirzajani, Hadi; Department of Electrical and Electronics Engineering; Koç University Research Center for Translational Medicine (KUTTAM) / Koç Üniversitesi Translasyonel Tıp Araştırma Merkezi (KUTTAM); College of EngineeringMicroneedle biosensors have emerged as a promising tool for in situ biomarker detection due to their minimally invasive nature and ability to interface with interstitial fluid (ISF). However, most previously demonstrated ones are limited to in situ detection of small molecules and ions, employing amperometry or potentiometry measurement techniques with electrical current or voltage output metrics, respectively, which may not be suitable for detecting large molecules, such as proteins. This letter presents an innovative approach utilizing a microneedle array integrated with an interdigitated electrode (MAIDE), enabling in situ capacitive detection and quantification of protein biomarkers. Following microneedle penetration, the interdigitated electrode array establishes direct contact with the solution, enabling real-time monitoring of interfacial capacitance modulations as the result of the binding reaction, leading to the acquisition of rich molecular data. Equivalent circuit model extraction followed by impedance spectroscopy for different concentrations of bovine serum albumin (BSA) indicated the suitability of the proposed platform in tracking the interfacial capacitance variations with respect to different BSA concentrations of 100, 10, and 1 μg/mL with a detection limit of 21 ng/mL. Furthermore, the device showed satisfactory results for biodegradability experiments where it disintegrated for a duration of 10 h. In addition, in vivo experiments show stable capacitance readings with (dC/C)% deviations less than 0.5%, indicating its potential for biodegradable wearable/implantable capacitive biosensing applicationsPublication Metadata only A wideband silicon photonic duplexer constructed from a deep photonic network of custom Mach-Zehnder interferometers(Society of Photographic Instrumentation Engineers (SPIE), 2024) Department of Electrical and Electronics Engineering; Amiri, Ali Najjar; Görgülü, Kazım; Mağden, Emir Salih; Department of Electrical and Electronics Engineering; Graduate School of Sciences and Engineering; College of EngineeringUsing a highly-scalable and physics-informed design platform with custom Mach-Zehnder interferometers (MZIs), we design and experimentally demonstrate a 1 x 2 wideband duplexer on silicon operating within 1450-1630 nm. The device is constructed from six layers of cascaded MZIs whose geometries are optimized using an equivalent artificial neural network, in a total timeframe of 75 seconds. Experimental results show below 0.72 dB deviation from the arbitrarily-specified target response, and less than 0.66 dB insertion loss. Demonstrated capabilities and the computational efficiency of our design framework pave the way towards the scalable deployment of custom MZI networks in communications, sensing, and computation applications.Publication Metadata only Experimental demonstration of a broadband, ultra-compact, and fabrication-tolerant silicon photonic 10% power tap(Society of Photographic Instrumentation Engineers (SPIE), 2024) Department of Electrical and Electronics Engineering; Daşdemir, Ahmet Onur; Mağden, Emir Salih; Department of Electrical and Electronics Engineering; Graduate School of Sciences and Engineering; College of EngineeringInverse design approaches with topology optimization can yield in highly efficient devices; however designing fabrication-compatible, broadband, yet simultaneously fabrication-tolerant devices still widely remains a challenge. Here, we design a broadband and fabrication-tolerant 10% silicon-based power tap using 3D-FDTD simulations and topology optimization, and demonstrate its experimental performance. The power tap has a compact footprint of 7.0 mu mx3.1 mu m, and achieves a broadband and spectrally flat operation from 1500 nm to 1600 nm. The device was specifically built to be fabrication-tolerant using an approach that maintains high performance under over-etch and under-etch scenarios by maximizing the contiguous area of the silicon layer in the final device. This tolerance was verified with 3D-FDTD simulations with 15 nm over-etch and under-etch modifications, demonstrating a change of less than 0.64 dB at either output port compared to the original device response at 1550 nm. The designed power tap was fabricated using a standard 220 nm thick silicon-on-insulator platform. The experimental measurements match closely with the design target and 3D-FDTD results, achieving state-of-the-art performance with excess losses as low as 0.23 dB and broadband operation. The output ports of the device also exhibit extremely flat spectra, where the transmission remains between 0.86 and 0.92 for the through port, and between 0.06 and 0.14 for the tap port throughout the 1500-1600 nm spectral range. These results represent the state-of-the-art experimental performance in compact power taps, and prove the effectiveness of fabrication-tolerant optimization.Publication Metadata only Machine learning-based PHY-authentication without prior attacker information for wireless multiple access channels(Springer, 2024) Department of Electrical and Electronics Engineering; Altun, Ufuk; Başar, Ertuğrul; Department of Electrical and Electronics Engineering; Graduate School of Sciences and Engineering; College of EngineeringPhysical layer (PHY) authentication methods provide spatial security by exploiting the unique channel between two users. In recent years, many studies focused on substituting traditional threshold-based detection mechanisms with machine/deep learning classifiers to solve the threshold selection problem and obtain better detection accuracy. However, these studies assume that receivers have access to spoofer's channel information at the training of the classifier, which is unrealistic for real-time scenarios. In this study, we propose a PHY-authentication architecture for wireless multiple access channels (W-MACs) that removes this assumption and works without any prior information about the spoofer. The proposed method is designed for multi-user systems and is suitable for any classifier model or communication protocol. The feasibility and the performance of the proposed method are investigated via computer simulations and compared with a benchmark model. The results proved the feasibility of the proposed method as it can detect spoofers successfully without requiring spoofers' channel information.