Constructing the dynamical law. by Robert Marsland (8616483)

Algoritmo de clasificación de expresiones de odio por tipos en español (Algorithm for classifying hate expressions by type in Spanish) by Daniel Pérez Palau (11097348)

“…</p><h2>Model Architecture</h2><p dir="ltr">The model is based on <code>pysentimiento/robertuito-base-uncased</code> with the following modifications:</p><ul><li>A dense classification layer was added over the base model</li><li>Uses input IDs and attention masks as inputs</li><li>Generates a multi-class classification with 5 hate categories</li></ul><h2>Dataset</h2><p dir="ltr"><b>HATEMEDIA Dataset</b>: Custom hate speech dataset with categorization by type:</p><ul><li><b>Labels</b>: 5 hate type categories (0-4)</li><li><b>Preprocessing</b>:</li><li>Null values ​​removed from text and labels</li><li>Reindexing and relabeling (original labels are adjusted by subtracting 1)</li><li>Exclusion of category 2 during training</li><li>Conversion of category 5 to category 2</li></ul><h2>Training Process</h2><h3>Configuration</h3><ul><li><b>Batch size</b>: 128</li><li><b>Epoches</b>: 5</li><li><b>Learning rate</b>: 2e-5 with 10% warmup steps</li><li><b>Early stopping</b> with patience=2</li><li><b>Class weights</b>: Balanced to handle class imbalance</li></ul><h3>Custom Metrics</h3><ul><li>Recall for specific classes (focus on class 2)</li><li>Precision for specific classes (focus on class 3)</li><li>F1-score (weighted)</li><li>AUC-PR</li><li>Recall at precision=0.6 (class 3)</li><li>Precision at recall=0.6 (class 2)</li></ul><h2>Evaluation Metrics</h2><p dir="ltr">The model is evaluated using:</p><ul><li>Macro recall, precision, and F1-score</li><li>One-vs-Rest AUC</li><li>Accuracy</li><li>Per-class metrics</li><li>Confusion matrix</li><li>Full classification report</li></ul><h2>Technical Features</h2><h3>Data Preprocessing</h3><ul><li><b>Tokenization</b>: Maximum length of 128 tokens (truncation and padding)</li><li><b>Encoding of labels</b>: One-hot encoding for multi-class classification</li><li><b>Data split</b>: 80% training, 10% validation, 10% testing</li></ul><h3>Optimization</h3><ul><li><b>Optimizer</b>: Adam with linear warmup scheduling</li><li><b>Loss function</b>: Categorical Crossentropy (from_logits=True)</li><li><b>Imbalance handling</b>: Class weights computed automatically</li></ul><h2>Requirements</h2><p dir="ltr">The following Python packages are required:</p><ul><li>TensorFlow</li><li>Transformers</li><li>scikit-learn</li><li>pandas</li><li>datasets</li><li>matplotlib</li><li>seaborn</li><li>numpy</li></ul><h2>Usage</h2><ol><li><b>Data format</b>:</li></ol><ul><li>CSV file or Pandas DataFrame</li><li>Required column name: <code>text</code> (string type)</li><li>Required column name: Data type label (integer type, 0-4) - optional for evaluation</li></ul><ol><li><b>Text preprocessing</b>:</li></ol><ul><li>Automatic tokenization with a maximum length of 128 tokens</li><li>Long texts will be automatically truncated</li><li>Handling of special characters, URLs, and emojis included</li></ul><ol><li><b>Label encoding</b>:</li></ol><ul><li>The model classifies hate speech into 5 categories (0-4)</li><li><code>0</code>: Political hatred: Expressions directed against individuals or groups based on political orientation.…”

Data_Sheet_1_Decoding emotional resilience in aging: unveiling the interplay between daily functioning and emotional health.PDF by Minhua Guo (18392646)

Antibody challenge outcomes. by Andrea Blasco (7439102)

“…Also shown is the benchmark algorithm implemented in Python (A1) and C++ (A2); note that benchmark algorithms A1 and A2 have perfect accuracy (<i>ACC</i> equal to unity). …”

BOFdat Step 3: Identifying species-specific metabolic end goals. by Jean-Christophe Lachance (6619307)

Comparison against the commercial scanner. by Swathi Lakshmi Balasubramanian (12078093)

SParse EXact (SPEX) LU and Cholesky Factorization Library by Erick Moreno-Centeno (19460626)

“…<p dir="ltr">Exact solutions of sparse systems of linear equations (SLEs) are of fundamental importance for some applications within mathematics (e.g., in computer-assisted proofs), computer science (e.g., in computing radial basis functions for scattered data interpolation), and engineering (e.g., in studies of anharmonic oscillations in semiconductors). …”

DataSheet1_Development of a Multilayer Deep Neural Network Model for Predicting Hourly River Water Temperature From Meteorological Data.docx by Reza Abdi (3636907)

“…We trained the LR and DNN algorithms on Google’s TensorFlow model using Keras artificial neural network library on Python. …”

<b>AI for imaging plant stress in invasive species </b>(dataset from the article https://doi.org/10.1093/aob/mcaf043) by Erola Fenollosa (20977421)

“…The described extracted features were used to predict leaf betalain content (µg per FW) using multiple machine learning regression algorithms (Linear regression, Ridge regression, Gradient boosting, Decision tree, Random forest and Support vector machine) using the <i>Scikit-learn</i> 1.2.1 library in Python (v.3.10.1) (list of hyperparameters used is given in <a href="#sup1" target="_blank">Supplementary Data S5</a>). …”

Predictive Analysis of Mushroom Toxicity Based Exclusively on Their Natural Habitat. by Enrico Bertozzi (22461709)

“…Multiple SVM models were trained and evaluated, including configurations with linear and RBF (Radial Basis Function) kernels. </p><p dir="ltr">Additionally, an exhaustive hyperparameter search was performed using GridSearchCV to optimize the C, gamma, and kernel parameters (testing 'linear,' 'rbf,' 'poly,' and 'sigmoid'), aiming to find the highest-performing configuration. …”

Code by Baoqiang Chen (21099509)

“…We implemented machine learning algorithms using the following R packages: rpart for Decision Trees, gbm for Gradient Boosting Machines (GBM), ranger for Random Forests, the glm function for Generalized Linear Models (GLM), and xgboost for Extreme Gradient Boosting (XGB). …”

Core data by Baoqiang Chen (21099509)

“…We implemented machine learning algorithms using the following R packages: rpart for Decision Trees, gbm for Gradient Boosting Machines (GBM), ranger for Random Forests, the glm function for Generalized Linear Models (GLM), and xgboost for Extreme Gradient Boosting (XGB). …”

MCCN Case Study 2 - Spatial projection via modelled data by Donald Hobern (21435904)

“…</p><h4><b>Data sources</b></h4><ul><li><b>BradGinns_SOIL2004_SoilData.csv</b> - Soil measurements from the University of Sydney Llara Campey farm site from 2004, corresponding to sites L1, L3 and L4 describing mid-depth, soil apparent electrical conductivity (ECa), GammaK, Clay, Silt, Sand, pH and soil electrical conductivity (EC)</li><li><b>Llara_Campey_field_boundaries_poly.shp</b> - Field boundary shapes for the University of Sydney Llara Campey farm site</li></ul><h4><b>Dependencies</b></h4><ul><li>This notebook requires Python 3.10 or higher</li><li>Install relevant Python libraries with: <b>pip install mccn-engine rocrate rioxarray pykrige</b></li><li>Installing mccn-engine will install other dependencies</li></ul><h4><b>Overview</b></h4><ol><li>Select soil sample measurements for pH or EC at 45 cm depth</li><li>Split sample measurements into 80% subset to model interpolated layers and 20% to test interpolated layers</li><li>Generate STAC metadata for layers</li><li>Load data cube</li><li>Interpolate pH and EC across site using the 80% subset and three different 2D interpolation methods from rioxarray (nearest, linear and cubic) and one from pykrige (linear)</li><li>Calculate the error between each layer of interpolated values and measured values for the 20% setaside for testing</li><li>Compare the mean and standard deviation of the errors for each interpolation method</li><li>Clean up and package results as RO-Crate</li></ol><h4><b>Notes</b></h4><ul><li>The granularity of variability in soil data significantly compromises all methods</li><li>Depending on the 80/20 split, different methods may appear more reliable, but the pykrige linear method is most often best</li></ul><p><br></p>…”

PySilsub—a toolbox for silent substitution by Joel Martin (11864048)

“…Device settings that will produce lights to selectively stimulate the photoreceptor(s) of interest can be found using a variety of analytic and algorithmic approaches. Here we present <em>PySilSub</em>, a novel Python package for silent substitution featuring object-oriented support for individual colorimetric observer models, multi-primary stimulation devices, and solving silent substitution problems with linear algebra and constrained numerical optimisation. …”

Landscape17 by Vlad Carare (22092515)

“…The NEB algorithm optimizes these interpolations of 20 images with a 50 eV/Å spring constant and convergence criterion of 10⁻² eV/Å for the maximum force component. …”

An Ecological Benchmark of Photo Editing Software: A Comparative Analysis of Local vs. Cloud Workflows by Pierre-Alexis DELAROCHE (22092572)

“…Performance Profiling Algorithms Energy Measurement Methodology # Pseudo-algorithmic representation of measurement protocol def capture_energy_metrics(workflow_type: WorkflowEnum, asset_vector: List[PhotoAsset]) -> EnergyProfile: baseline_power = sample_idle_power_draw(duration=30) with PowerMonitoringContext() as pmc: start_timestamp = rdtsc() # Read time-stamp counter if workflow_type == WorkflowEnum.LOCAL: result = execute_local_pipeline(asset_vector) elif workflow_type == WorkflowEnum.CLOUD: result = execute_cloud_pipeline(asset_vector) end_timestamp = rdtsc() energy_profile = EnergyProfile( duration=cycles_to_seconds(end_timestamp - start_timestamp), peak_power=pmc.get_peak_consumption(), average_power=pmc.get_mean_consumption(), total_energy=integrate_power_curve(pmc.get_power_trace()) ) return energy_profile Statistical Analysis Framework Our analytical pipeline employs advanced statistical methodologies including: Variance Decomposition: ANOVA with nested factors for hardware configuration effects Regression Analysis: Generalized Linear Models (GLM) with log-link functions for energy modeling Temporal Analysis: Fourier transform-based frequency domain analysis of power consumption patterns Cluster Analysis: K-means clustering with Euclidean distance metrics for workflow classification Data Validation and Quality Assurance Measurement Uncertainty Quantification All energy measurements incorporate systematic and random error propagation analysis: Instrument Precision: ±0.1W for CPU power, ±0.5W for GPU power Temporal Resolution: 1ms sampling with Nyquist frequency considerations Calibration Protocol: NIST-traceable power standards with periodic recalibration Environmental Controls: Temperature-compensated measurements in climate-controlled facility Outlier Detection Algorithms Statistical outliers are identified using the Interquartile Range (IQR) method with Tukey's fence criteria (Q₁ - 1.5×IQR, Q₃ + 1.5×IQR). …”