威普罗

Air pollution is responsible for various health issues, including respiratory and cardiovascular diseases, among individuals. However, previous studies have not successfully identified the sources of air pollution that contribute to the acceleration of climate change. To address this gap, a novel approach known as the Adaptive Exponential Sigmoid Fuzzy Tsallis Entropy Interference System (AES-FTEIS) is proposed for identifying air pollution sources. Datasets from remote sensing and ground-level air pollution measurements are collected, temporally aligned using the Prior Distribution Regularized Kalman Filter (PDRKF), and imputed using Cross-Entropy Minimization Spline Interpolation (CEMSI). Additionally, aerosol particles such as PM2.5 and PM10 are extracted from the dataset and incorporated into the analysis. Subsequently, the data are organized by location and time using Transfer Entropy Spectral Clustering (TESC), and their correlation are analysed using Spearman Rank Correlation (SRC). An exploratory data analysis is conducted on the time-based grouped results through Box plots, leading to feature extraction. Finally, the AES-FTEIS is utilized to identify the pollution sources based on the levels of aerosol particle concentrations. The experimental results show that the proposed WOLSTM-ASLRCNN classifier achieves 97.56% of accuracy and 98.5% of precision, outperforming existing models such as CNN (95.3%), LSTM (93.84%), GRU (92.57%), and RNN (89.66%). The proposed TESC clustering method obtained a silhouette score of 0.9721, higher than SC (0.9687), AC (0.9428), HC (0.9271) and KMC (0.9087). Moreover, the AES-FTEIS source identification approach reduced the rule generation time to 1483 ms, demonstrating the effectiveness of the proposed framework.

When humans translate, not every word depends equally on the surrounding context. Some tokens, particularly function words like pronouns and auxiliaries, rely heavily on preceding or following sentences, while others, such as proper nouns, do not. Understanding this inherent context sensitivity is essential for evaluating whether machine translation systems use context in human-like ways. However, existing approaches to analysing context usage rely on discourse-specific test sets or model internals, making them narrow or model-dependent. We propose a post-hoc, model-agnostic framework to quantify context sensitivity at lexical and syntactic levels using two measures derived from word alignments: fertility (number of target tokens generated per source token) and entropy (stability of fertility patterns across contexts). Using reference translations for three language pairs (German ↔ English, English → Hindi) under four context conditions, we show that context selectively redistributes generative responsibility from source to context tokens without altering overall fertility. Function words show the largest fertility reductions, while content words remain stable, suggesting that context resolves ambiguity rather than adding new information. Our framework provides a ground-truth characterisation of selective context usage in human translation, establishing a diagnostic baseline for evaluating machine translation models.

The Robotic manipulation in dynamic environments, stable and collision-free manipulation of objects is the most important consideration. In this work, a novel approach is proposed that combines two robust techniques—Dexterity Networks (Dex-Net) for grasping objects and Rapidly Exploring Random Trees (RRT) for path planning. Dex-Net uses an enormous 3D model database to make guesses at stable grasps, and this allows robots to grasp wide numbers of objects, even potentially cluttered objects. RRT then augments this by efficiently searching for safe paths in high-dimensional environments. Together, the algorithms allow for robots to manipulate objects independently within dense environments with confidence. The combined system excels in comparison with isolated methods with 94.7

A comparative energy consumption analysis of AMD EPYC–based and Intel Xeon based head nodes in an AI appliance with 8 GPUs is presented. A detailed energy model is developed covering the head node’s CPU, memory, I/O, storage, and cooling power, as well as the GPUs’ compute, memory, and I/O power. Using empirical data from an 8xGPU system (NVIDIA H100 GPUs) and vendor documentation, the two CPU platforms are compared in terms of power draw and efficiency. Both configurations reach a similar peak power (approximately 8–9 kW for 8 GPUs fully loaded) dominated by the GPUs. The AMD EPYC head node (high core count, high memory bandwidth) better feeds the GPUs. Energy efficiency depends on workload: for heavy GPU bound training for large CNN or Transformer models with big batches, both AMD and Intel head nodes contribute < 15% of total system power (only a few hundred watts), making GPU efficiency the primary factor. In more CPU intensive or I/O heavy workloads (e.g. recommender systems with CPU hosted embeddings), AMD’s greater core count and memory throughput improve performance per watt, whereas Intel’s platform may idle at lower baseline power for lighter loads. Key energy-saving strategies are also summarized: optimizing batch sizes and concurrency (which in one case cut training energy by 4×) [1], leveraging power management features (AMD Infinity power tuning, Intel optimized power modes), and component tuning (CPU DVFS, GPU power capping, etc.). These findings provide guidance for data center practitioners to maximize GPU utilization and overall efficiency in multi-GPU AI servers.

ABSTRAK ABSTRAK Keputusan pembelian dalam masyarakat kontemporer telah bergeser dari memenuhi kebutuhan fungsional menuju konsumsi simbolik. Penelitian ini bertujuan untuk mendekonstruksi faktor-faktor determinan yang membentuk perilaku konsumen, dengan fokus pada tiga variabel utama: determinasi sosiologis, tingkat pendapatan, dan hegemoni iklan. Menggunakan metode deskriptif-analitis dan pendekatan fenomenologi yang diperkaya oleh pengalaman empiris peneliti sebagai praktisi riset pasar, penelitian ini mengungkap bahwa rasionalitas ekonomi sering kali dikalahkan oleh tekanan sosial dan manipulasi psikologis. Hasil penelitian menunjukkan bahwa: (1) Secara sosiologis, pembelian didominasi oleh faktor konformitas lingkungan dan kekuatan habituasi (kebiasaan) yang menciptakan inersia kognitif; (2) Dari sisi ekonomi, peningkatan pendapatan tidak selalu berbanding lurus dengan inflasi gaya hidup, melainkan memicu rasionalitas fungsional pada segmen masyarakat yang mapan; dan (3) Iklan berfungsi menciptakan "kebutuhan buatan" (kebutuhan buatan) serta menumbuhkan kesadaran merek (brand awareness) yang menjadi pencipta mutlak terjadinya transaksi. Kesimpulannya, pola konsumsi adalah hasil interseksi antara ketersediaan informasi (iklan), validasi sosial (gengsi), dan justifikasi ekonomi. Kata Kunci: Perilaku Konsumen, Sosiologi Ekonomi, Habituasi, Hegemoni Iklan, Daya Beli. ABSTRAK Keputusan pembelian dalam masyarakat kontemporer telah bergeser dari pemenuhan fungsional ke konsumsi simbolis. Studi ini bertujuan untuk menguraikan faktor-faktor penentu yang membentuk perilaku konsumen, dengan fokus pada tiga variabel utama: determinasi sosiologis, tingkat pendapatan, dan hegemoni periklanan. Dengan menggunakan metode deskriptif-analitis dan pendekatan fenomenologis yang diperkaya oleh pengalaman empiris peneliti sebagai praktisi riset pasar, studi ini mengungkapkan bahwa rasionalitas ekonomi seringkali dikalahkan oleh tekanan sosial dan manipulasi psikologis. Temuan menunjukkan bahwa: (1) Secara sosiologis, pembelian didominasi oleh kesesuaian lingkungan dan kekuatan pembiasaan, yang menciptakan inersia kognitif; (2) Secara ekonomi, peningkatan pendapatan tidak selalu berkorelasi linier dengan inflasi gaya hidup tetapi lebih memicu rasionalitas fungsional di segmen masyarakat yang mapan; dan (3) Iklan berfungsi untuk menciptakan "kebutuhan buatan" dan menanamkan kesadaran merek sebagai prasyarat mutlak untuk transaksi. Kesimpulannya, pola konsumsi adalah hasil dari persimpangan antara ketersediaan informasi (iklan), validasi sosial (prestise), dan pembenaran ekonomi. Kata kunci: Perilaku Konsumen, Sosiologi Ekonomi, Habituasi, Hegemoni Periklanan, Daya Beli. 📢 BACA VERSI LENGKAP & DISKUSI INTERAKTIF: Ingin membaca analisis ini dengan bahasa yang lebih ringan dan studi kasus nyata? Kunjungi artikel selengkapnya di Blog Resmi KunciPro: 👉 [https://www.sosiolegal.com/] KunciPro Research Institute - Membongkar Kebenaran, Melawan Arus.