Multiple Choice Question Answering (MCQA) is an important problem with numerous real-world applications, such as medicine, law, and education. The high cost of building MCQA datasets makes few-shot learning pivotal in this domain. While Large Language Models (LLMs) can enable few-shot learning, their direct application in real-world scenarios is often hindered by their high computational cost. To address this challenge, we propose a simple yet effective approach that uses LLMs for data generation and scoring. Our approach utilizes LLMs to create MCQA data which contains questions and choices, and to assign probability scores to the generated choices. We then use the generated data and LLM-assigned scores to finetune a smaller and more efficient encoder-only model, DeBERTa-v3-base by leveraging distillation loss. Extensive experiments on the Massive Multitask Language Understanding (MMLU) benchmark demonstrate that our method improves accuracy from 28.9% to 39.3%, representing a gain of over 10% compared to a baseline finetuned directly on 5-shot examples. This shows the effectiveness of LLM-driven data generation and knowledge distillation for few-shot MCQA.

Analysing malware is important to understand how malicious software works and to develop appropriate detection and prevention methods. Dynamic analysis can overcome evasion techniques commonly used to bypass static analysis and provide insights into malware runtime activities. Much research on dynamic analysis focused on investigating machine-level information (e.g., CPU, memory, network usage) to identify whether a machine is running malicious activities. A malicious machine does not necessarily mean all running processes on the machine are also malicious. If we can isolate the malicious process instead of isolating the whole machine, we could kill the malicious process, and the machine can keep doing its job. Another challenge dynamic malware detection research faces is that the samples are executed in one machine without any background applications running. It is unrealistic as a computer typically runs many benign (background) applications when a malware incident happens. Our experiment with machine-level data shows that the existence of background applications decreases previous state-of-the-art accuracy by about 20.12% on average. We also proposed a process-level Recurrent Neural Network (RNN)-based detection model. Our proposed model performs better than the machine-level detection model; 0.049 increase in detection rate and a false-positive rate below 0.1.

Last July, photographer Reuben Wu and a crew of around 30 people hiked from the Peruvian city of Huaraz, nestled in the Cordillera Blanca region of the Andes, to the 16,000-foot-high Pastoruri glacier. The hike took around four hours and the crew arrived after sunset, finding the melting glacier lit only by a full moon. These Stitched Photos of Greenland's Icebergs Are Sew Great Wu has shot conceptual landscape photography in some of the world's most remote locations--East Java, Patagonia, Chile's Atacama Desert, Norway's Svalbard Archipelago--but this shoot, part of a mini-documentary about Wu's photography done as part of a Coors Light ad campaign, gave him the opportunity to highlight global warming by photographing a fast-receding glacier, one of the last in South America. "There were parts of the glacier where you could see evidence of pretty extreme breakdown and melting of the snow," Wu says. "Parts of the glacier no longer had the epic, jagged chunks of ice."

PetaBencana brings together local flood information and mobile mapping for Jakarta, Bandung and Surabaya, Indonesia - allowing millions of citizens to share real-time flood information. The platform has recently added AWS Rekognition to assist emergency operators make more informed decision on flood information shared by citizens.