arXiv subjects

Md Millat Hosen

Publications and source records attributed to Md Millat Hosen.

2 recordsLinked to original sources

arXiv2026

Push-Pull Determinants Among Bangladeshi Students Enrolled in NCR Private Universities: A Single-Destination Exploratory Study

International student mobility from Bangladesh is a significant feature of South Asian higher education, yet India's National Capital Region (NCR) remains underexplored as a destination for outbound Bangladeshi students. This exploratory single-destination study examines push-pull factors among Bangladeshi students enrolled at private universities in India's NCR. A structured online survey was administered to students at Sharda University, Noida International University, and Galgotias University (n = 63; n = 56 after quality filtering). Descriptive statistics, K-means clustering, and binary logistic regression were applied within Lee's push-pull framework. Political and administrative disruption in Bangladesh's academic calendar was the leading push factor (M = 3.73). Geographical and cultural proximity was the strongest pull factor (M = 3.80), closely followed by visa accessibility (M = 3.73), with no significant mean difference between them. Australia and Germany were the most frequently considered alternative destinations (33.9% each). Advisory networks influenced 73.2% of respondents under a broad threshold and 66.1% under a stricter threshold, mainly for university and course selection. Satisfaction and recommendation intent were positively associated, and infrastructure satisfaction showed the strongest association with high recommendation intent in a five-predictor logistic model (OR = 2.54, 95% CI [1.09, 5.89]). K-means clustering produced three exploratory decision-profile groups: Comprehensively Motivated, Proximity-Led Enrollers, and Low-Salience Enrollers. The findings suggest that geographic nearness, cultural familiarity, and visa accessibility may operate as a composite accessibility advantage in this short-haul intra-regional corridor.

cs.CY↗

arXiv2026

Fine-Tuning a 7B Advisor on Free-Tier GPUs: An Adapter-Handoff Recipe and a Synthetic-Data Reliability Caution

Fine-tuning a 7B language model for specialized advising is attractive in resource-constrained settings, but multi-epoch runs routinely exceed the wall-clock limits of the free-tier GPUs (Kaggle, Colab) such users rely on. We report two things. First, a practical recipe: a three-epoch QLoRA fine-tune of Mistral-7B-Instruct-v0.3 (4-bit NF4, LoRA rank 16, via Unsloth) completed across two free-tier 16 GB GPUs (Tesla P100 then T4) by checkpointing only the small LoRA adapter (41.9M parameters) and resuming on the second machine. Adapter-only handoff is sufficient -- optimizer and scheduler state need not be transferred -- so the binding constraint is per-step VRAM and per-session wall-clock, not aggregate compute. Second, and more importantly, an honest evaluation that returns a cautionary result. On a blind held-out comparison against the un-fine-tuned base model, the fine-tuned model scored higher on similarity to the synthetic training distribution (BERTScore F1 +0.063, a fidelity not quality signal) but lower on advising quality: a blind LLM-as-judge preferred the base model on 46% of prompts versus 18%, and a source-verified factuality audit found four confident errors from the fine-tuned model on policy-sensitive topics against zero for the base. Auditing the training data with the same method, we find this is not a fine-tuning artifact: each audited error is already present in the Gemini-generated training answers, and a random-sample audit finds verifiable errors in a sizable fraction of responses (28-40%; single-judge, n=40). The data is therefore sufficient to account for the errors, which we attribute to the synthetic-data pipeline rather than the adapter-handoff method. We release the dataset, adapter, cross-GPU notebooks, and full evaluation harness so every result reproduces on a single 16 GB GPU.

cs.AI↗