Publications without Fulltext
Permanent URI for this collectionhttps://hdl.handle.net/20.500.14288/3
Browse
695 results
Filters
Advanced Search
Filter by
Settings
Search Results
Item Metadata only Precise event sampling-based data locality tools for AMD multicore architectures(Wiley, 2023) 0000-0002-6166-4252; 0000-0002-2351-0770; Chabbi, Milind; Kelly, Paul H.; Department of Computer Engineering; Department of Computer Engineering; Sasongko, Muhammad Aditya; Erten, Didem Unat; Researcher; Faculty Member; College of Engineering; College of Engineering; N/A; 219274We propose ComDetective+, an inter-thread communication analyzer, and ReuseTracker+, a reuse distance analyzer, that leverage the hardware features in AMD processors to support low-overhead profiling. Both tools employ the instruction-based sampling (IBS) facility and debug registers in AMD processors to detect inter-thread communication and data reuse. Different from prior arts, ComDetective+ differentiates the communication into true and false sharing, and ReuseTracker+ measures reuse distance in private and shared caches by also considering cache line invalidation with low overhead. Both tools can attribute the communications and reuses to source code lines. To our knowledge these tools are two of the few profiling tools designed specifically for AMD x86 architectures using IBS. Our tools are timely and relevant considering the rise in numbers of AMD processor based data centers and HPC systems. We perform experiments to evaluate the accuracy and overheads of the proposed tools on an AMD machine with two-socket EPYC 7352 processors. ComDetective+ exhibits high accuracy while introducing 5.14xruntime and 1.4x memory overheads. ReuseTracker+ also displays high accuracy, which is 95%, with 11.76x runtime and 1.46x memory overheads. These overheads are much lower than the overheads of existing simulators and code instrumentation-based tools. Lastly, we demonstrate the usage of the tools by having COMDETECTIVE+ and REUSETRACKER+ facilitate the code refactoring of two data mining benchmarks to improve their performance by up to 29%.Item Metadata only Precise event sampling on AMD versus intel: quantitative and qualitative comparison(IEEE Computer Soc, 2023) 0000-0002-6166-4252; 0000-0002-2351-0770; Chabbi, Milind; Kelly, Paul H. J.; Department of Computer Engineering; Department of Computer Engineering; Sasongko, Muhammad Aditya; Erten, Didem Unat; Researcher; Faculty Member; College of Engineering; College of Engineering; N/A; 219274Precise event sampling is a profiling feature in commodity processors that can sample hardware events and accurately locate the instructions that trigger the events. This feature has been used in a large number of tools to detect application performance issues. Although precise event sampling is readily supported in modern multicore architectures, vendor supports exhibit great differences that affect their accuracy, stability, overhead, and functionality. This work presents the most comprehensive study to date on benchmarking the event sampling features of Intel PEBS and AMD IBS and performs in-depth analysis on key differences through series of microbenchmarks. Our qualitative and quantitative analysis shows that PEBS allows finer-grained and more accurate sampling of hardware events, while IBS offers richer set of information at each sample though it suffers from lower accuracy and stability. Moreover, OS signal delivery, which is a common method used by the profiling software, introduces significant time overhead to the original overhead incurred by the hardware mechanisms in both PEBS and IBS. We also found that both PEBS and IBS have bias in sampling events across multiple different locations in a code. Lastly, we demonstrate how our findings on microbenchmarks under different thread counts hold for a full-fledged profiling tool that runs on the state-of-the-art Intel and AMD machines. Overall our detailed comparisons serve as a great reference and provide invaluable information for hardware designers and profiling tool developers.Item Metadata only Efficacy and tolerability of immediate switch from sodium channel blockers to Lacosamide(Academic Press Inc Elsevier Science, 2023) 0000-0002-3752-1825; 0000-0001-7718-4299; N/A; 0000-0002-7671-7097; N/A; Yilmaz, Melek Kandemir; Atmaca, Murat Mert; Guler, Selda Keskin; N/A; N/A; N/A; Department of Computer Engineering; N/A; Gürses, Rabia Candan; Çelebi, Özlem; Buluş, Eser; Duman Arda; Özgün, Orhan Talha; Faculty Member; Doctor; Doctor; Undergraduate Student; Undergraduate Student; School of Medicine; N/A; N/A; College of Engineering; School of Medicine; Koç University Hospital; 110149; N/A; N/A; N/A; N/ALacosamide (LCM) is a new-generation anti-seizure medication approved for monotherapy and add-on therapy for focal-onset epilepsy. It has novel pharmacodynamics and favorable pharmacokinetic qualities with good clinical response. This study aims to evaluate the effectiveness and tolerability of LCM when used in the immediate switch from sodium channel blockers in patients with focal-onset and generalized-onset epilepsies. This retrospective, multicenter observational study was conducted with adult patients who received LCM as mono-or polytherapy through immediate switch with 6 to 52 months follow-up. The clinical data obtained during the follow-up period were analyzed to assess retention rate, seizure freedom, more than 50% seizure reduction, and adverse effects. A total of 32 patients (eight females, 24 males) with a median age of 49.75 (range, 23-86) years, median age at epilepsy onset of 32.58 (range, 0.5-85) years, and median epilepsy duration of 17.17 (range, 1-46) years were included in this study. Seizure frequency was between 1 and 90 in the past 6 months. Seven (21.9%) of the patients had structural brain lesions and 27 (84.4%) of the patients had EEG abnormalities. The adverse effects leading to switching were hyponatremia, rash, elevated liver enzymes, pain, and erectile dysfunction. At 14.34 (range, 6-52) months follow-up, 30 (93.75%) patients in total retained LCM, 20 (66.7%) of them were seizure-free, and 13 were on LCM monotherapy. Responder rate was 81.25%. Eight (25%) of the patients experienced adverse effects after the immediate switch. One patient with generalized-onset epi-lepsy needed to quit LCM due to an increase in seizures. Seizure frequency did not change in three patients in the focal-onset group. Immediate switch to LCM showed favorable outcomes with a signifi-cant reduction in seizure frequency, high retention rates, and tolerable adverse effect profiles in both focal-onset and generalized-onset seizures.& COPY; 2023 Elsevier Inc. All rights reserved.Item Metadata only Perception-distortion trade-off in the SR space spanned by flowmodels(IEEE, 2022) 0000-0003-1465-8121; N/A; 0000-0002-5078-4590; 0000-0002-6280-8422; Erdem, Erkut; Department of Electrical and Electronics Engineering; N/A; Department of Electrical and Electronics Engineering; Department of Computer Engineering; Tekalp, Ahmet Murat; Korkmaz, Cansu; Doğan, Zafer; Erdem, Aykut; Faculty Member; PhD Student; Faculty Member; Faculty Member; Koç Üniversitesi İş Bankası Yapay Zeka Uygulama ve Araştırma Merkezi (KUIS AI)/ Koç University İş Bank Artificial Intelligence Center (KUIS AI); College of Engineering; Graduate School of Sciences and Engineering; College of Engineering; College of Engineering; 26207; N/A; 280658; 20331Flow-based generative super-resolution (SR) models learn to produce a diverse set of feasible SR solutions, called the SR space. Diversity of SR solutions increases with the temperature (t) of latent variables, which introduces random variations of texture among sample solutions, resulting in visual artifacts and low fidelity. In this paper, we present a simple but effective image ensembling/fusion approach to obtain a single SR image eliminating random artifacts and improving fidelity without significantly compromising perceptual quality. We achieve this by benefiting from a diverse set of feasible photorealistic solutions in the SR space spanned by flow models. We propose different image ensembling and fusion strategies which offer multiple paths to move sample solutions in the SR space to more desired destinations in the perception-distortion plane in a controllable manner depending on the fidelity vs. perceptual quality requirements of the task at hand. Experimental results demonstrate that our image ensembling/fusion strategy achieves more promising perception-distortion trade-off compared to sample SR images produced by flow models and adversarially trained models in terms of both quantitative metrics and visual quality.Item Metadata only Structural coverage of the human interactome(Oxford University Press, 2023) 0000-0002-2297-2113; 0000-0002-4202-4049; 0000-0002-0389-9459; 0000-0002-0349-4312; 0009-0000-4377-8865; Department of Computer Engineering; Department of Chemical and Biological Engineering; Department of Chemical and Biological Engineering; N/A; N/A; Gürsoy, Attila; Keskin, Özlem; Tunçbağ, Nurcan; Kösoğlu, Kayra; Aydın, Zeynep; Faculty Member; Faculty Member; Faculty Member; PhD Student; PhD Student; College of Engineering; College of Engineering; College of Engineering; Graduate School of Sciences and Engineering; Graduate School of Sciences and Engineering; 8745; 26605; 245513; N/A; N/AComplex biological processes in cells are embedded in the interactome, representing the complete set of protein–protein interactions. Mapping and analyzing the protein structures are essential to fully comprehending these processes’ molecular details. Therefore, knowing the structural coverage of the interactome is important to show the current limitations. Structural modeling of protein–protein interactions requires accurate protein structures. In this study, we mapped all experimental structures to the reference human proteome. Later, we found the enrichment in structural coverage when complementary methods such as homology modeling and deep learning (AlphaFold) were included. We then collected the interactions from the literature and databases to form the reference human interactome, resulting in 117 897 non-redundant interactions. When we analyzed the structural coverage of the interactome, we found that the number of experimentally determined protein complex structures is scarce, corresponding to 3.95% of all binary interactions. We also analyzed known and modeled structures to potentially construct the structural interactome with a docking method. Our analysis showed that 12.97% of the interactions from HuRI and 73.62% and 32.94% from the filtered versions of STRING and HIPPIE could potentially be modeled with high structural coverage or accuracy, respectively. Overall, this paper provides an overview of the current state of structural coverage of the human proteome and interactome. © 2024 Oxford University Press. All rights reserved.Item Metadata only Use of affective visual information for summarization of human-centric videos(Institute of Electrical and Electronics Engineers Inc., 2023) 0000-0002-2715-2368; 0000-0003-2238-137X; Department of Computer Engineering; N/A; Erzin, Engin; Köprü, Berkay; Faculty Member; PhD Student; Koç Üniversitesi İş Bankası Yapay Zeka Uygulama ve Araştırma Merkezi (KUIS AI)/ Koç University İş Bank Artificial Intelligence Center (KUIS AI); College of Engineering; Graduate School of Sciences and Engineering; 34503; N/AThe increasing volume of user-generated human-centric video content and its applications, such as video retrieval and browsing, require compact representations addressed by the video summarization literature. Current supervised studies formulate video summarization as a sequence-to-sequence learning problem, and the existing solutions often neglect the surge of the human-centric view, which inherently contains affective content. In this study, we investigate the affective-information enriched supervised video summarization task for human-centric videos. First, we train a visual input-driven state-of-the-art continuous emotion recognition model (CER-NET) on the RECOLA dataset to estimate activation and valence attributes. Then, we integrate the estimated emotional attributes and their high-level embeddings from the CER-NET with the visual information to define the proposed affective video summarization (AVSUM) architectures. In addition, we investigate the use of attention to improve the AVSUM architectures and propose two new architectures based on temporal attention (TA-AVSUM-GRU) and spatial attention (SA-AVSUM-GRU). We conduct video summarization experiments on the TvSum and COGNIMUSE datasets. The proposed temporal attention-based TA-AVSUM architecture attains competitive video summarization performances with strong improvements for the human-centric videos compared to the state-of-the-art in terms of F-score, self-defined face recall, and rank correlation metrics. © 2010-2012 IEEE.Item Metadata only On the effectiveness of re-identification attacks and local differential privacy-based solutions for smart meter data(Scitepress, 2023) 0000-0002-7676-0167; N/A; Department of Computer Engineering; Department of Computer Engineering; Gürsoy, Mehmet Emre; Kaya, Zeynep Sıla; Faculty Member; Undergraduate Student; College of Engineering; College of Engineering; 330368; N/ASmart meters are increasing the ability to collect, store and share households' energy consumption data. On the other hand, the availability of such data raises novel privacy concerns. Although the data can be de-identified or pseudonymized, a critical question remains: How unique are households' energy consumptions, and is it possible to re-identify households based on partial or imperfect knowledge of their consumption? In this paper, we aim to answer this question, and make two main contributions. First, we develop an adversary model in which an adversary who observes a pseudonymized dataset and knows a limited number of consumption readings from a target household aims to infer which record in the dataset corresponds to the target. We characterize the adversary's knowledge by two parameters: number of known readings and precision of readings. Using experiments conducted on three real-world datasets, we demonstrate that the adversary can indeed achieve high inference rates. Second, we propose a local differential privacy (LDP) based solution for protecting the privacy of energy consumption data. We evaluate the impact of our LDP solution on three datasets using two utility metrics, three LDP protocols, and various parameter settings. Results show that our solution can attain high accuracy and low estimation error under strong privacy guarantees.Item Metadata only Multi-GPU communication schemes for iterative solvers: when CPUs are not in charge(Association for Computing Machinery, 2023) 0000-0002-2351-0770; 0000-0002-9603-2466; 0000-0001-7235-6418; N/A; Wahib, Mohamed; Department of Computer Engineering; N/A; N/A; N/A; Erten, Didem Unat; Sağbili Doğan; Baydamirli Javid; Ismayilov, Ismayil; Faculty Member; PhD Student; PhD Student; Master Student; College of Engineering; Graduate School of Sciences and Engineering; Graduate School of Sciences and Engineering; Graduate School of Sciences and Engineering; 219274; N/A; N/A; N/AThis paper proposes a fully autonomous execution model for multi-GPU applications that completely excludes the involvement of the CPU beyond the initial kernel launch. In a typical multi-GPU application, the host serves as the orchestrator of execution by directly launching kernels, issuing communication calls, and acting as a synchronizer for devices. We argue that this orchestration, or control flow path, causes undue overhead and can be delegated entirely to devices to improve performance in applications that require communication among peers. For the proposed CPU-free execution model, we leverage existing techniques such as persistent kernels, thread block specialization, device-side barriers, and device-initiated communication routines to write fully autonomous multi-GPU code and achieve significantly reduced communication overheads. We demonstrate our proposed model on two broadly used iterative solvers, 2D/3D Jacobi stencil and Conjugate Gradient(CG). Compared to the CPU-controlled baselines, the CPU-free model can improve 3D stencil communication latency by 58.8% and provide a 1.63x speedup for CG on 8 NVIDIA A100 GPUs. The project code is available at https://github.com/ParCoreLab/CPU-Free-model. © 2023 Owner/Author(s).Item Metadata only An overview of affective speech synthesis and conversion in the deep learning era(IEEE-Inst Electrical Electronics Engineers Inc, 2023) 0000-0002-1524-1646; N/A; Triantafyllopoulos, Andreas; Schuller, Bjorn W.; He, Xiangheng; Yang, Zijiang; Tzirakis, Panagiotis; Liu, Shuo; Mertes, Silvan; Andre, Elisabeth; Fu, Ruibo; Tao, Jianhua; Department of Computer Engineering; N/A; Sezgin, Tevfik Metin; İymen Gökçe; Faculty Member; Master Student; Koç Üniversitesi İş Bankası Yapay Zeka Uygulama ve Araştırma Merkezi (KUIS AI)/ Koç University İş Bank Artificial Intelligence Center (KUIS AI); College of Engineering; Graduate School of Sciences and Engineering; 18632; N/ASpeech is the fundamental mode of human communication, and its synthesis has long been a core priority in human-computer interaction research. In recent years, machines have managed to master the art of generating speech that is understandable by humans. However, the linguistic content of an utterance encompasses only a part of its meaning. Affect, or expressivity, has the capacity to turn speech into a medium capable of conveying intimate thoughts, feelings, and emotions-aspects that are essential for engaging and naturalistic interpersonal communication. While the goal of imparting expressivity to synthesized utterances has so far remained elusive, following recent advances in text-to-speech synthesis, a paradigm shift is well under way in the fields of affective speech synthesis and conversion as well. Deep learning, as the technology that underlies most of the recent advances in artificial intelligence, is spearheading these efforts. In this overview, we outline ongoing trends and summarize state-of-the-art approaches in an attempt to provide a broad overview of this exciting field.Item Metadata only Corrigendum to "Associations between sleep characteristics and glycemic variability in youth with type 1 diabetes" [Sleep Med. 109 (2023) 132-142](Elsevier B.V., 2023) 0000-0002-2614-0832; 0000-0003-3781-3892; 0000-0001-6312-6004; 0000-0002-8889-6811; 0000-0002-7855-1297; 0000-0003-3919-7763; 0000-0003-1633-9570; Boran, Perran; Barış, Hatice Ezgi; Us, Mahmut Caner; Aygün, Burcu; Haliloğlu, Belma; Bereket, Abdullah; N/A; N/A; N/A; N/A; Department of Computer Engineering; N/A; N/A; İpar, Necla; Gökçe, Tuğba; Can, Ecem; Eviz, Elif; İnan, Neslihan Gökmen; Mutlu, Rahime Gül Yeşiltepe; Hatun, Şükrü; Doctor; Doctor; Nurse; Researcher; Teaching Faculty; Faculty Member; Faculty Member; N/A; N/A; N/A; School of Medicine; College of Engineering; School of Medicine; School of Medicine; Koç University Hospital; N/A; N/A; N/A; 327618; 285581; 153511; 153504[The authors regret < that there was an error in the prevalence of T1DM in the introduction section. The correct prevalence should be 0.75/1000 in Turkey>. The authors would like to apologise for any inconvenience caused. © 2023 Elsevier B.V.