Search

Multi-aspect controllable text generation aims to control text generation in attributes from multiple aspects, making it a complex but powerful task in natural language processing. Supervised fine-tuning methods are often employed for this task due to their simplicity and effectiveness. However, they still have some limitations: low rank adaptation (LoRA) only fine-tunes a few parameters and has suboptimal control effects, while full fine-tuning (FFT) requires significant computational resources and is susceptible to overfitting, particularly when data is limited. Moreover, existing works typically train multi-aspect controllable text generation models using only single-aspect annotated data, which results in discrepancies in data distribution; at the same time, accurately generating text with specific attributes is a challenge that requires strong attribute-aware capabilities. To address these limitations, we propose a lightweight, adaptive and attribute-aware framework for multi-aspect controllable text generation. Our framework can dynamically adjust model parameters according to different aspects of data to achieve controllable text generation, aiming to optimize performance across multiple aspects. Experimental results show that our framework outperforms other strong baselines, achieves state-of-the-art performance, adapts well to data discrepancies, and is more accurate in attribute perception., Comment: 17 pages,9 figures

Large Language Models (LLMs) have achieved remarkable progress in recent years; however, their excellent performance is still largely limited to major world languages, primarily English. Many LLMs continue to face challenges with multilingual tasks, especially when it comes to low-resource languages. To address this issue, we introduced Marco-LLM: Massive multilingual training for cross-lingual enhancement LLM. We have collected a substantial amount of multilingual data for several low-resource languages and conducted extensive continual pre-training using the Qwen2 models. This effort has resulted in a multilingual LLM named Marco-LLM. Through comprehensive evaluations on various multilingual benchmarks, including MMMLU, AGIEval, Belebele, Flores-200, XCOPA and many others, Marco-LLM has demonstrated substantial improvements over state-of-the-art LLMs. Furthermore, Marco-LLM achieved substantial enhancements in any-to-any machine translation tasks, showing the effectiveness of our multilingual LLM. Marco-LLM is a pioneering multilingual LLM designed to not only perform exceptionally well in multilingual tasks, including low-resource languages, but also maintain strong performance in English and other major languages, closing the performance gap between high- and low-resource language capabilities. By bridging languages, this effort demonstrates our dedication to ensuring LLMs work accurately across various languages.

The energy efficiency and power of a quantum thermoelectric system with multiple electric currents and only one heat currents are studied. The system is connected to the hot heat bath with one terminal but the cold bath with multiple terminals or vice versal. We find that the cooperative effects can be a potentially useful tool in improving the energy efficiency and output power in multi-terminal mesoscopic thermoelectric systems. As an example, we show that the cooperation between the two thermoelectric effects in three-terminal thermoelectric systems leads to markedly improved performance of heat engine within the linear response regime using the Landauer-B\"{u}tiker formalism. Such improvement also emerge in four-terminal thermoelectric heat engines with three output electric currents. Cooperative effects in these multi-terminal thermoelectric systems can significantly enlarge the physical parameter region with high efficiency and power. For refrigeration, we find that the energy efficiency can also be substantially improved if multi-terminal configurations are considered, suggesting a useful scheme toward electronic cooling. Our study illustrates cooperative effects as a convenient approach toward high-performance thermoelectric energy conversion in multi-terminal mesoscopic systems., Comment: 12 pages, 8 figures

We establish a theory of optimal efficiency and power for three-terminal thermoelectric engines which have two independent output electric currents and one input heat current. This set-up goes beyond the conventional heat engines with only one output electric current. For such a set-up, we derive the optimal efficiency and power and their trade-off for three-terminal heat engines with and without time-reversal symmetry. The formalism goes beyond the known optimal efficiency and power for systems with or without time-reversal symmetry, showing interesting features that have not been revealed before. A concrete example of quantum-dot heat engine is studied to show that the current set-up can have much improved efficiency and power compared with previous set-ups with only one output electric current. Our analytical results also apply for thermoelectric heat engines with multiple output electric currents, providing an alternative scheme toward future high-performance thermoelectric materials.

The energy efficiency and power of a three-terminal thermoelectric nanodevice are studied by considering elastic tunneling through a single quantum dot. Facilitated by the three-terminal geometry, the nanodevice is able to generate simultaneously two electrical powers by utilizing only one temperature bias. These two electrical powers can add up constructively or destructively, depending on their signs. It is demonstrated that the constructive addition leads to the enhancement of both energy efficiency and output power for various system parameters. In fact, such enhancement, dubbed as thermoelectric cooperative effect, can lead to maximum efficiency and power no less than when only one of the electrical power is harvested.