GATech at AbjadGenEval Shared Task: Multilingual Embeddings for Arabic Machine-Generated Text Classification

Ahmed Khaled Khamis

cs.CLFeb 17, 2026

GATech at AbjadGenEval Shared Task: Multilingual Embeddings for Arabic Machine-Generated Text Classification

Authors: Ahmed Khaled Khamis

Organizations: Georgia Institute of Technology

Abstract

We present our approach to the AbjadGenEval shared task on detecting AI-generated Arabic text. We fine-tuned the multilingual E5-large encoder for binary classification, and we explored several pooling strategies to pool token representations, including weighted layer pooling, multi-head attention pooling, and gated fusion. Interestingly, none of these outperformed simple mean pooling, which achieved an F1 of 0.75 on the test set. We believe this is because complex pooling methods introduce additional parameters that need more data to train properly, whereas mean pooling offers a stable baseline that generalizes well even with limited examples. We also observe a clear pattern in the data: human-written texts tend to be significantly longer than machine-generated ones.

Figures & tables

Statistic	Human	Machine
Samples	2,649	2,649
Avg. Words	632.0	303.0
Avg. Characters	3,806.7	1,865.5
Max Words	3,068	1,969

Table 1: Dataset statistics by class. Human-written texts are significantly longer than machine-generated texts.

Parameter	Value
Model	multilingual-e5-large
Max Sequence Length	512 tokens
Batch Size	16
Gradient Accumulation	4 steps
Effective Batch Size	64
Learning Rate	$2\times 10^{-5}$

Table 2: Training hyperparameters.

Pooling Strategy	Test F1
Mean Pooling	0.75
Weighted Layer Pooling + Attention
+ Gated Fusion	0.70
Weighted Layer Pooling + Attention	0.71

Table 3: Pooling strategy comparison on development set (2 samples). All methods achieved perfect dev scores, but mean pooling performed best on test.

Figure 1: Dataset analysis: (a) Balanced class distribution with 2,649 samples per class. (b) Word count distribution showing human texts are significantly longer.

Multilingual Text Classification

See all 20 papers

GATech at AbjadMed: Bidirectional Encoders vs. Causal Decoders: Insights from 82-Class Arabic Medical Classification
Feb 17, 2026Ahmed Khaled KhamisArabic NLP Clinical NLP
Georgia Institute of Technology
Fine-Tuning Pre-Trained Code Models for AI-Generated Code Detection
May 2, 2026Jany-Gabriel Ispas, Sergiu Nisioi
Human Language Technologies Research Center Faculty of Mathematics and Computer Science University of Bucharest
Spotlights and Blindspots: Evaluating Machine-Generated Text Detection
Apr 17, 2026Kevin Stowe, Kailash PatilAI-Generated Text Detection Language Model Generation Evaluation
Pindrop

Feb 17, 2026J/K move · Enter open · S save

GATech at AbjadMed: Bidirectional Encoders vs. Causal Decoders: Insights from 82-Class Arabic Medical Classification

Ahmed Khaled Khamis

Georgia Institute of Technology

Arabic NLP Clinical NLP Decoder-Only Language Models

Open paper AskPDF arXiv

This paper presents system description for Arabic medical text classification across 82 distinct categories. Our primary architecture utilizes a fine-tuned AraBERTv2 encoder enhanced with a hybrid pooling strategies, combining attention and mean representations, and multi-sample dropout for robust regularization. We systematically benchmark this approach against a suite of multilingual and Arabic-specific encoders, as well as several large-scale causal decoders, including zero-shot re-ranking via Llama 3.3 70B and feature extraction from Qwen 3B hidden states. Our findings demonstrate that specialized bidirectional encoders significantly outperform causal decoders in capturing the precise semantic boundaries required for fine-grained medical text classification. We show that causal decoders, optimized for next-token prediction, produce sequence-biased embeddings that are less effective for categorization compared to the global context captured by bidirectional attention. Despite significant class imbalance and label noise identified within the training data, our results highlight the superior semantic compression of fine-tuned encoders for specialized Arabic NLP tasks. Final performance metrics on the test set, including Accuracy and Macro-F1, are reported and discussed.

Abstract

Multilingual Text Classification

Explore similar work

GATech at AbjadMed: Bidirectional Encoders vs. Causal Decoders: Insights from 82-Class Arabic Medical Classification

Fine-Tuning Pre-Trained Code Models for AI-Generated Code Detection

Spotlights and Blindspots: Evaluating Machine-Generated Text Detection