Long-term agent memory is increasingly multimodal, yet existing evaluations rarely test whether agents preserve the visual evidence needed for later reasoning. In prior work, many visually grounded questions can be answered using only captions or textual traces, allowing answers to be inferred without preserving the fine-grained visual evidence. Meanwhile, harder cases that require reasoning over changing visual states are largely absent. Therefore, we introduce MemEye, a framework that evaluates memory capabilities from two dimensions: one measures the granularity of decisive visual evidence (from scene-level to pixel-level evidence), and the other measures how retrieved evidence must be used (from single evidence to evolutionary synthesis). Under this framework, we construct a new benchmark across 8 life-scenario tasks, with ablation-driven validation gates for assessing answerability, shortcut resistance, visual necessity, and reasoning structure. By evaluating 13 memory methods across 4 VLM backbones, we show that current architectures still struggle to preserve fine-grained visual details and reason about state changes over time. Our findings show that long-term multimodal memory depends on evidence routing, temporal tracking, and detail extraction.
Large foundation models have made significant advances in embodied intelligence, enabling synthesis and reasoning over egocentric input for household tasks. However, VLM-based auto-labeling is often noisy because the primary data sources lack accurate human action labels, chain-of-thought (CoT), and spatial annotations; these errors are amplified during long-horizon spatial instruction following. These issues stem from insufficient coverage of minute-long, daily household planning tasks and from inaccurate spatial grounding. As a result, VLM reasoning chains and world-model synthesis can hallucinate objects, skip steps, or fail to respect real-world physical attributes. To address these gaps, we introduce EgoTL. EgoTL builds a think-aloud capture pipeline for egocentric data. It uses a say-before-act protocol to record step-by-step goals and spoken reasoning with word-level timestamps, then calibrates physical properties with metric-scale spatial estimators, a memory-bank walkthrough for scene context, and clip-level tags for navigation instructions and detailed manipulation actions. With EgoTL, we are able to benchmark VLMs and World Models on six task dimensions from three layers and long-horizon generation over minute-long sequences across over 100 daily household tasks. We find that foundation models still fall short as egocentric assistants or open-world simulators. Finally, we finetune foundation models with human CoT aligned with metric labels on the training split of EgoTL, which improves long-horizon planning and reasoning, step-wise reasoning, instruction following, and spatial grounding.
Micro, Small, and Medium Enterprises (MSMEs) play an important role in Indonesia's economy, but because the quality of their financial reporting is still low, it becomes one of the obstacles in business decision-making and access to financing. This study aims to analyze the effect of accounting competence and financial literacy on the quality of MSME financial reports, as well as to test the role of digital application adoption as a moderating variable. The research approach used is quantitative with a survey method. Data was collected through a Likert-scale questionnaire distributed to 125 MSME actors in Medan Amplas District. Data analysis was conducted using the Structural Equation Modeling – Partial Least Squares (SEM-PLS) method. The research is expected to show that accounting competence and financial literacy have a positive impact on the quality of financial reports, and that adopting digital applications strengthens this relationship. This study is at technology readiness level (TRL) 3. It aligns with SDG 4 – Strengthening Knowledge and Literacy. Improving the quality of financial reports will increase access to financing and investment for SMEs, in line with SDG 8 – Decent Work and Economic Growth. Digitalizing financial reports is a form of innovation that enhances the competitiveness of SMEs. Developing this infrastructure aligns with SDG 9 – Industry, Innovation, and Infrastructure. This research supports Asta Cita-6 of the President of Indonesia, regarding (Strengthening the Empowerment of SMEs and Cooperatives), which aims to boost economic independence and the competitiveness of national SMEs in a sustainable way.
Penelitian ini bertujuan untuk menganalisis hubungan antara limit PayLater dan kepemilikan akun e-commerce dengan frekuensi belanja online mahasiswa, serta keterkaitannya dengan pengelolaan keuangan pribadi. Latar belakang penelitian ini berangkat dari meningkatnya penggunaan layanan keuangan digital di kalangan mahasiswa, yang berpotensi memengaruhi perilaku konsumsi dan pengelolaan keuangan. Metode yang digunakan adalah pendekatan kuantitatif dengan uji korelasi Spearman Rank, karena data tidak berdistribusi normal. Hasil penelitian menunjukkan bahwa tidak terdapat hubungan yang signifikan antara limit PayLater dengan frekuensi belanja online (ρ = 0,230; p = 0,138) maupun antara jumlah akun e-commerce dengan frekuensi belanja online (ρ = 0,133; p = 0,395). Temuan ini menunjukkan bahwa besarnya limit PayLater dan banyaknya akun e-commerce tidak secara langsung memengaruhi intensitas belanja mahasiswa. Hal ini mengindikasikan bahwa faktor lain, seperti kontrol diri dan kemampuan pengelolaan keuangan pribadi, kemungkinan berperan lebih besar dalam membentuk perilaku konsumtif mahasiswa di era digital.
This research demonstrates the importance of identifying and developing food crop agribusiness, particularly lowland rice, in Indonesia and Malaysia. It is hoped that this research will formulate a model for developing food crop agribusiness in Indonesia and Malaysia. The research subjects were 60 respondents of rice paddy farmers consisting of 30 respondents of rice paddy farmers in Indonesia and 30 respondents of rice paddy farmers in Malaysia and registered in farmer groups, which became the object of research, namely the appropriate strategies and models in developing food crop agribusiness. Data Analysis Method To analyze the problems studied using descriptive analysis, namely providing clear and in-depth problem information interpreted according to the results of the research conducted based on theoretical support related to the research object to measure farmer perceptions by analyzing using the Likert scale method. Farmer groups in the research area not only serve as a forum for improving the quality and quantity of production, but also as a channel for introducing the latest technology, expanding market access, and improving farmer welfare through collective marketing and economic empowerment. From the Index Score obtained, there is a comparison of agricultural institutional models between Indonesia and Malaysia. From the benchmark of farmer group institutional activity, the farming subsystem (on-farm), Indonesia is higher than Malaysia. From the benchmark of production facility subsystem (Upstream), marketing & processing system (Downstream), supporting service subsystem, Malaysia is higher than Indonesia, while inhibiting factors and potential for farmer institutional development between Indonesia and Malaysia still need to be improved again so that farmers are more prosperous. In general, Malaysian agricultural institutions are more supportive in the development of food crop agribusiness than Indonesia with a score difference of 3.9%. With an index score of 73.4% higher than Indonesia's 69.5%, it shows the adoption of agricultural technology and more equitable irrigation infrastructure that ensures a more stable water supply, so that the average yield per hectare is maximized. Higher levels of use of modern agricultural mechanization (such as combine harvesters and monitoring drones) and large-scale commercial land management are more dominant in Malaysia. Farmer groups in the research area not only serve as a forum for improving the quality and quantity of production, but also as a channel for introducing the latest technology, expanding market access, and improving farmer welfare through collective marketing and economic empowerment. From the Index Score obtained, there is a comparison of agricultural institutional models between Indonesia and Malaysia. From the benchmark of farmer group institutional activity, the farming subsystem (on-farm), Indonesia is higher than Malaysia. From the benchmark of production facility subsystem (Upstream), marketing & processing system (Downstream), supporting service subsystem, Malaysia is higher than Indonesia, while inhibiting factors and potential for farmer institutional development between Indonesia and Malaysia still need to be improved again so that farmers are more prosperous. In general, Malaysian agricultural institutions are more supportive in the development of food crop agribusiness than Indonesia with a score difference of 3.9%. With an index score of 73.4% higher than Indonesia's 69.5%, it shows the adoption of agricultural technology and more equitable irrigation infrastructure that ensures a more stable water supply, so that the average yield per hectare is maximized. Higher levels of use of modern agricultural mechanization (such as combine harvesters and monitoring drones) and large-scale commercial land management are more dominant in Malaysia.
