Doyoung Kim

Hyeonbin Hwang, Doyoung Kim, Seungone Kim et al. · cmu

Training on large amounts of rationales (i.e., CoT Fine-tuning) is effective at improving the reasoning capabilities of large language models (LLMs). However, acquiring human-authored rationales or augmenting rationales from proprietary models is costly and not scalable. In this paper, we study the problem of whether LLMs could self-improve their reasoning capabilities. To this end, we propose Self-Explore, where the LLM is tasked to explore the first wrong step (i.e., the first pit) within the rationale and use such signals as fine-grained rewards for further improvement. On the GSM8K and MATH test set, Self-Explore achieves 11.57% and 2.89% improvement on average across three LLMs compared to supervised fine-tuning (SFT). Our code is available at https://github.com/hbin0701/Self-Explore.

Tyler Griggs, Xiaoxuan Liu, Jiaxiang Yu et al.

Large language models (LLMs) are increasingly integrated into many online services, yet they remain cost-prohibitive to deploy due to the requirement of expensive GPU instances. Prior work has addressed the high cost of LLM serving by improving the inference engine, but less attention has been given to selecting the most cost-efficient GPU type(s) for a specific LLM service. There is a large and growing landscape of GPU types and, within these options, higher cost does not always lead to increased performance. Instead, through a comprehensive investigation, we find that three key LLM service characteristics (request size, request rate, SLO) strongly influence GPU cost efficiency, and differing GPU types are most cost efficient for differing LLM service settings. As a result, the most cost-efficient allocation for a given service is typically a mix of heterogeneous GPU types. Based on this analysis, we introduce Mélange, a GPU allocation framework that navigates these diverse LLM service characteristics and heterogeneous GPU option space to automatically and efficiently derive the minimal-cost GPU allocation for a given LLM service. We formulate the GPU allocation task as a cost-aware bin packing problem where GPUs are bins and items are slices of the service workload. Our formulation's constraints account for a service's unique characteristics, allowing Mélange to be flexible to support diverse service settings and heterogeneity-aware to adapt the GPU allocation to a specific service. Compared to using only a single GPU type, Mélange reduces deployment costs by up to 77% in conversational settings, 33% in document-based settings, and 51% in a mixed setting.

Doyoung Kim, Jongwon Lee, Jinho Park et al.

Language models' ability to extrapolate learned behaviors to novel, more complex environments beyond their training scope is highly unknown. This study introduces a path planning task in a textualized Gridworld to probe language models' extrapolation capabilities. We show that conventional approaches, including next token prediction and Chain of Thought (CoT) finetuning, fail to extrapolate in larger, unseen environments. Inspired by human cognition and dual process theory, we propose cognitive maps for path planning, a novel CoT framework that simulates humanlike mental representations. Our experiments show that cognitive maps not only enhance extrapolation to unseen environments but also exhibit humanlike characteristics through structured mental simulation and rapid adaptation. Our finding that these cognitive maps require specialized training schemes and cannot be induced through simple prompting opens up important questions about developing general-purpose cognitive maps in language models. Our comparison with exploration-based methods further illuminates the complementary strengths of offline planning and online exploration.

Kyongsik Yun, Luan Nguyen, Tuan Nguyen et al.

It is important to find the target as soon as possible for search and rescue operations. Surveillance camera systems and unmanned aerial vehicles (UAVs) are used to support search and rescue. Automatic object detection is important because a person cannot monitor multiple surveillance screens simultaneously for 24 hours. Also, the object is often too small to be recognized by the human eye on the surveillance screen. This study used UAVs around the Port of Houston and fixed surveillance cameras to build an automatic target detection system that supports the US Coast Guard (USCG) to help find targets (e.g., person overboard). We combined image segmentation, enhancement, and convolution neural networks to reduce detection time to detect small targets. We compared the performance between the auto-detection system and the human eye. Our system detected the target within 8 seconds, but the human eye detected the target within 25 seconds. Our systems also used synthetic data generation and data augmentation techniques to improve target detection accuracy. This solution may help the search and rescue operations of the first responders in a timely manner.