Aligning Large Language Models (LLMs) with diverse and potentially conflicting human values necessitates navigating complex multi-objective landscapes.However, existing prompt-conditioned approaches face a critical training-inference discrepancy: they rely on ground-truth scores during training while requiring manual user-specification at inference.We introduce prediction of implicit preferences to bridge this gap while reducing user burden.To this end, we propose Self-Guided Alignment (SGA), a framework that transforms passive reward dependency into an intrinsic adaptive sensing capability.It employs a dual-head architecture to unify preference internalization with conditional generation, enabling the model to learn a latent mapping between raw prompts and preference profiles.Through adaptive preference sensing, the model autonomously predicts the latent preference score to self-guide the generation, thereby eliminating the need for manual specification at inference.Extensive experiments across diverse model scales demonstrate that SGA often outperforms state-of-the-art baselines, achieving superior multi-objective trade-offs and improved preference alignment.

More and more recent studies highlight the crucial role of the human microbiome in maintaining health, while modern advancements in metagenomic sequencing technologies have been accumulating data that are associated with human diseases. Although metagenomic data offer rich, multifaceted information, including taxonomic and functional abundance profiles, their full potential remains underutilized, as most approaches rely only on one type of information to discover and understand their related correlations with respect to disease occurrences. To address this limitation, we propose a multistage fusion tabular transformer architecture (MSFT-Transformer), aiming to effectively integrate various types of high-dimensional tabular information extracted from metagenomic data. Its multistage fusion strategy consists of three modules: a fusion-aware feature extraction module in the early stage to improve the extracted information from inputs, an alignment-enhanced fusion module in the mid stage to enforce the retainment of desired information in cross-modal learning, and an integrated feature decision layer in the late stage to incorporate desired cross-modal information. We conduct extensive experiments to evaluate the performance of MSFT-Transformer over state-of-the-art models on five standard datasets. Our results indicate that MSFT-Transformer provides stable performance gains with reduced computational costs. An ablation study illustrates the contributions of all three models compared with a reference multistage fusion transformer without these novel strategies. The result analysis implies the significant potential of the proposed model in future disease prediction with metagenomic data.

Tabular datasets commonly have high dimensionality and limited sample sizes in biomedical field. In such contexts, neural networks often struggle with overfitting, leading to weaker performance when compared to linear and tree-based models. However, despite the abundance of auxiliary domain information on biomedical data, it is often overlooked, even though it has the potential to significantly improve neural network performance. To address this issue, we propose a Knowledge-Infused Network with Diverse Sparsity and In-Context Learning (KNDI), a novel deep learning architecture designed to handle high-dimensional, small-sample genomic data. KNDI consists of a novel Diverse Sparsity Feature Abstraction (DSFA) module and a Locally-Aligned Contextual Learning (LACL) module to infuse auxiliary biomedical knowledge and boost the performance. The DSFA module extracts diverse high-level features from high-dimensional inputs by integrating a sparsity mechanism to reduce dimensionality and applying contrastive constraints to achieve feature diversity. Subsequently, the LACL module adaptively learns contextual information from high-level feature embeddings, incorporating locality-aware interactions and preserving the original information through self-alignment, facilitating subsequent multiplication fusion. Extensive experiments show that KNDI achieves an average Pearson correlation coefficient improvement of 10.45% compared to the best performing baselines on six high-dimensional, low-sample-size genomic datasets.