Selected work

Representative Papers

PRISM: A Predictive Protocol for Permutation Optimization via Landscape Diagnostics

Aug 08, 2026

This work addresses the challenge of efficiently determining whether permutation optimization is worthwhile—and which search strategy to employ—in scenarios where system components are fixed but their ordering significantly impacts performance. The authors propose PRISM, a novel protocol that introduces fitness landscape diagnostics into permutation optimization for the first time. By leveraging low-cost first-order autocorrelation and fitness-distance correlation analyses, PRISM predicts search behavior prior to optimization, thereby guiding the selection of appropriate search strategies, operators, and simplification schemes. Empirical validation across diverse tasks—including neural architecture design, scientific machine learning, and large-model instruction sequencing—demonstrates that PRISM accurately forecasts optimization outcomes and confirms the complementary nature of permutation and content optimization.

0 citationsRead paper

Towards Digital Preservation of Efik: TTS for a Low-Resource African Language

Jul 05, 2026

This study addresses the challenges of speech synthesis and digital preservation for Efik, a low-resource African tonal language. The authors present the first end-to-end text-to-speech (TTS) system for Efik, leveraging a newly collected 3-hour single-speaker speech corpus. They establish reproducible TTS baselines using VITS, MMS-TTS, SpeechT5, and Orpheus-TTS, and evaluate performance through subjective assessments including MOS, Nat-MOS, and A-MOS. Among the models, MMS-TTS achieves the highest quality (MOS: 3.80 ± 0.63) and demonstrates greater stability in synthesizing long utterances, though it still exhibits tonal inaccuracies. This work provides the first systematic evaluation framework for TTS in low-resource tonal languages and underscores the need for larger-scale corpora and improved tonal modeling to advance synthesis quality.

0 citationsRead paper

WebChallenger: A Reliable and Efficient Generalist Web Agent

Jun 09, 2026

This work addresses the high reasoning cost, poor reusability, and lack of human-like cognitive mechanisms in general-purpose large language model agents for autonomous web navigation, particularly their inefficiency in repetitive tasks. The authors propose a model-size-agnostic agent framework that emulates human cognitive strengths through architectural design: it introduces PageMem, a DOM-based structured page representation, and integrates three key mechanisms—divide-and-conquer observation (mimicking selective attention), a lightweight exploration memory system (enabling structured memory), and a composite action workflow (enhancing operational fluency). Without fine-tuning any open-source large language models, the approach achieves success rates of 56.3%, 48.7%, 51.0%, and 70.9% on WebArena, VisualWebArena, Online-Mind2Web, and WorkArena, respectively, matching the performance of state-of-the-art closed-source systems while significantly reducing computational costs.

0 citationsRead paper

Sample-Size Scaling of the African Languages NLI Evaluation

Jun 02, 2026

This study addresses the scarcity of annotated data for natural language inference (NLI) in African languages and investigates whether increased data volume consistently improves performance. Conducting controlled scaling experiments on the AfriXNLI benchmark, the authors evaluate XLM-R Large and AfroXLM-R Large across 16 African languages using sample sizes ranging from 50 to 500, with multiple random subsampling runs to assess robustness. The findings reveal a non-monotonic, highly language-dependent relationship between sample size and NLI performance: for some languages, performance plateaus or even declines with more data, and variance remains substantial in low-resource settings. These results challenge the conventional assumption that “more data is always better,” suggesting that merely expanding labeled datasets may not reliably enhance NLI performance for African languages.

0 citationsRead paper

From Script to Semantics: Prompting Strategies for African NLI

Jun 02, 2026

This study addresses the underperformance of large language models on natural language inference tasks for low-resource African languages—such as Swahili, Yoruba, and Hausa—particularly in zero-shot prompting scenarios where systematic investigation has been lacking. The authors propose a language-aware prompt structure and conduct a systematic evaluation of five prompting strategies (Baseline, Script-Aware, Language-Specific, Contrastive, and NL-STP) using Llama3.2-3B and Gemma3-4B. Experimental results demonstrate that the proposed approach significantly improves overall accuracy and class balance across multilingual and multi-model settings. It not only outperforms existing zero-shot prompting methods but also surpasses strong baselines such as few-shot learning and chain-of-thought prompting, effectively mitigating the issue of class collapse.

0 citationsRead paper

Recent publications

Latest Papers

PRISM: A Predictive Protocol for Permutation Optimization via Landscape Diagnostics

Aug 08, 2026

This work addresses the challenge of efficiently determining whether permutation optimization is worthwhile—and which search strategy to employ—in scenarios where system components are fixed but their ordering significantly impacts performance. The authors propose PRISM, a novel protocol that introduces fitness landscape diagnostics into permutation optimization for the first time. By leveraging low-cost first-order autocorrelation and fitness-distance correlation analyses, PRISM predicts search behavior prior to optimization, thereby guiding the selection of appropriate search strategies, operators, and simplification schemes. Empirical validation across diverse tasks—including neural architecture design, scientific machine learning, and large-model instruction sequencing—demonstrates that PRISM accurately forecasts optimization outcomes and confirms the complementary nature of permutation and content optimization.

0 citationsRead paper

Towards Digital Preservation of Efik: TTS for a Low-Resource African Language

Jul 05, 2026

This study addresses the challenges of speech synthesis and digital preservation for Efik, a low-resource African tonal language. The authors present the first end-to-end text-to-speech (TTS) system for Efik, leveraging a newly collected 3-hour single-speaker speech corpus. They establish reproducible TTS baselines using VITS, MMS-TTS, SpeechT5, and Orpheus-TTS, and evaluate performance through subjective assessments including MOS, Nat-MOS, and A-MOS. Among the models, MMS-TTS achieves the highest quality (MOS: 3.80 ± 0.63) and demonstrates greater stability in synthesizing long utterances, though it still exhibits tonal inaccuracies. This work provides the first systematic evaluation framework for TTS in low-resource tonal languages and underscores the need for larger-scale corpora and improved tonal modeling to advance synthesis quality.

0 citationsRead paper

WebChallenger: A Reliable and Efficient Generalist Web Agent

Jun 09, 2026

This work addresses the high reasoning cost, poor reusability, and lack of human-like cognitive mechanisms in general-purpose large language model agents for autonomous web navigation, particularly their inefficiency in repetitive tasks. The authors propose a model-size-agnostic agent framework that emulates human cognitive strengths through architectural design: it introduces PageMem, a DOM-based structured page representation, and integrates three key mechanisms—divide-and-conquer observation (mimicking selective attention), a lightweight exploration memory system (enabling structured memory), and a composite action workflow (enhancing operational fluency). Without fine-tuning any open-source large language models, the approach achieves success rates of 56.3%, 48.7%, 51.0%, and 70.9% on WebArena, VisualWebArena, Online-Mind2Web, and WorkArena, respectively, matching the performance of state-of-the-art closed-source systems while significantly reducing computational costs.

0 citationsRead paper

Sample-Size Scaling of the African Languages NLI Evaluation

Jun 02, 2026

This study addresses the scarcity of annotated data for natural language inference (NLI) in African languages and investigates whether increased data volume consistently improves performance. Conducting controlled scaling experiments on the AfriXNLI benchmark, the authors evaluate XLM-R Large and AfroXLM-R Large across 16 African languages using sample sizes ranging from 50 to 500, with multiple random subsampling runs to assess robustness. The findings reveal a non-monotonic, highly language-dependent relationship between sample size and NLI performance: for some languages, performance plateaus or even declines with more data, and variance remains substantial in low-resource settings. These results challenge the conventional assumption that “more data is always better,” suggesting that merely expanding labeled datasets may not reliably enhance NLI performance for African languages.

0 citationsRead paper

From Script to Semantics: Prompting Strategies for African NLI

Jun 02, 2026

This study addresses the underperformance of large language models on natural language inference tasks for low-resource African languages—such as Swahili, Yoruba, and Hausa—particularly in zero-shot prompting scenarios where systematic investigation has been lacking. The authors propose a language-aware prompt structure and conduct a systematic evaluation of five prompting strategies (Baseline, Script-Aware, Language-Specific, Contrastive, and NL-STP) using Llama3.2-3B and Gemma3-4B. Experimental results demonstrate that the proposed approach significantly improves overall accuracy and class balance across multilingual and multi-model settings. It not only outperforms existing zero-shot prompting methods but also surpasses strong baselines such as few-shot learning and chain-of-thought prompting, effectively mitigating the issue of class collapse.

0 citationsRead paper