Predicting how software is used, shared, and evolves across publications is essential to studying scientific progress. Existing methods for representing software usage in publications rely mainly on tabular or textual formats, which limit their structural expressiveness and consequently their ability to predict software usage. We address these gaps by representing software mentions and citations as a graph and formulating software usage prediction as a link prediction task. To support this study, we construct the first large-scale graph dataset of publication and software mentions, Graph-TempCZ, covering 1959-2022 with over six million mention relationships. Experiments using both traditional machine learning and Graph Neural Network (GNN) show that graph-based models substantially outperform feature-based baselines, achieving a 5.98% improvement in test accuracy. Temporal experiments further reveal that models trained on one year generalize effectively to nearby years but show gradual performance decay as the temporal gap increases. This work provides the first comprehensive foundation for analyzing software usage through a temporal graph representation.

Thinking Mode Fusion (TMF) enables large language models to support both concise responses and long-form reasoning by unifying a non-thinking mode and a thinking mode within a single model. However, its training dynamics, including the data ratio and training schedule between the two modes, remain underexplored. In this work, we present a systematic study of TMF by analyzing the effects of the training schedule and data ratio between thinking and non-thinking modes. Focusing on mathematical problem solving, we construct a benchmark with multiple thinking-to-non-thinking data ratios and three training schedules. Our results reveal an asymmetric interaction between the two modes: increasing the ratio of non-thinking supervision reduces the accuracy of the thinking mode. We further show that different training schedules modulate this trade-off and that the optimal schedule depends on the data ratio. Finally, we quantify a negative correlation between non-thinking and thinking mode supervision, highlighting an inherent tension between these two modes. These findings provide practical guidance for designing effective TMF training settings. All code and data are released to support further research at: Fusion Bench.

Chemical molecules can be represented as graphs or as language descriptions. Training unimodal models on graphs results in different encodings than training them on language. Therefore, the existing literature force-aligns the unimodal models during training to use them in downstream applications such as drug discovery. But to what extent are graph and language unimodal model representations inherently aligned, i.e., aligned prior to any force-alignment training? Knowing this is useful for a more expedient and effective forced-alignment. For the first time, we explore methods to gauge the alignment of graph and language unimodal models. We find compelling differences between models and their ability to represent slight structural differences without force-alignment. We also present an unified unimodal alignment (U2A) benchmark for gauging the inherent alignment between graph and language encoders which we make available with this paper.

Existing parameter-efficient fine-tuning (PEFT) methods primarily fall into two categories: addition-based and selective in-situ adaptation. The former, such as LoRA, introduce additional modules to adapt the model to downstream tasks, offering strong memory efficiency. However, their representational capacity is often limited, making them less suitable for fine-grained adaptation. In contrast, the latter directly fine-tunes a carefully chosen subset of the original model parameters, allowing for more precise and effective adaptation, but at the cost of significantly increased memory consumption.To reconcile this trade-off, we propose NeuroAda, a novel PEFT method that enables fine-grained model finetuning while maintaining high memory efficiency. Our approach first identifies important parameters (i.e., connections within the network) as in selective adaptation, and then introduces bypass connections for these selected parameters. During finetuning, only the bypass connections are updated, leaving the original model parameters frozen.Empirical results on 23+ tasks spanning both natural language generation and understanding demonstrate that NeuroAda achieves state-of-the-art performance with as little as ≤ 0.02% trainable parameters, while reducing CUDA memory usage by up to 60%.We release our code here: https://github.com/FightingFighting/NeuroAda.git.