Recent studies have shown that post-aligning multilingual pretrained language models (mPLMs) using alignment objectives on both original and transliterated data can improve crosslingual alignment. This improvement further leads to better crosslingual transfer performance. However, it remains unclear how and why a better crosslingual alignment is achieved, as this technique only involves transliterations, and does not use any parallel data. This paper attempts to explicitly evaluate the crosslingual alignment and identify the key elements in transliteration-based approaches that contribute to better performance. For this, we train multiple models under varying setups for two pairs of related languages: (1) Polish and Ukrainian and (2) Hindi and Urdu. To assess alignment, we define four types of similarities based on sentence representations. Our experimental results show that adding transliterations alone improves the overall similarities, even for random sentence pairs. With the help of auxiliary transliteration-based alignment objectives, especially the contrastive objective, the model learns to distinguish matched from random pairs, leading to better crosslingual alignment. However, we also show that better alignment does not always yield better downstream performance, suggesting that further research is needed to clarify the connection between alignment and performance. The code implementation is based on \url{https://github.com/cisnlp/Transliteration-PPA}.

Large language models (LLMs) provide a promising way for accurate session-based recommendation (SBR), but they demand substantial computational time and memory. Knowledge distillation (KD)-based methods can alleviate these issues by transferring the knowledge to a small student, which trains a student based on the predictions of a cumbersome teacher. However, these methods encounter difficulties for LLM-based KD in SBR . 1) It is expensive to make LLMs predict for all instances in KD. 2) LLMs may make ineffective predictions for some instances in KD, e.g., incorrect predictions for hard instances or similar predictions as existing rec-ommenders for easy instances. In this paper, we propose an active LLM-based KD method in SBR, contributing to sustainable AI. To efficiently distill knowledge from LLMs with limited cost, we propose to extract a small proportion of instances predicted by LLMs. Meanwhile, for a more effective distillation, we propose an active learning strategy to extract instances that are as effective as possible for KD from a theoretical view. Specifically, we first formulate gains based on potential effects (e.g., effective, similar, and incorrect predictions by LLMs) and difficulties (e.g., easy or hard to fit) of instances for KD. Then, we propose to maximize the minimal gains of distillation to find the optimal selection policy for active learning, which can largely avoid extracting ineffective instances in KD. Experiments on real-world datasets show that our method significantly outperforms state-of-the-art methods for SBR. Introduction Recently, large language models (LLMs) have shown the potential to equip recommender systems (RSs) with their extensive knowledge and powerful reasoning capabilities (Zhao et al. 2023).

Recent advancements in visual generative models have enabled high-quality image and video generation, opening diverse applications. However, evaluating these models often demands sampling hundreds or thousands of images or videos, making the process computationally expensive, especially for diffusion-based models with inherently slow sampling. Moreover, existing evaluation methods rely on rigid pipelines that overlook specific user needs and provide numerical results without clear explanations. In contrast, humans can quickly form impressions of a model's capabilities by observing only a few samples. To mimic this, we propose the Evaluation Agent framework, which employs human-like strategies for efficient, dynamic, multi-round evaluations using only a few samples per round, while offering detailed, user-tailored analyses. It offers four key advantages: 1) efficiency, 2) promptable evaluation tailored to diverse user needs, 3) explainability beyond single numerical scores, and 4) scalability across various models and tools. Experiments show that Evaluation Agent reduces evaluation time to 10% of traditional methods while delivering comparable results. The Evaluation Agent framework is fully open-sourced to advance research in visual generative models and their efficient evaluation.

Constructing compact and informative 3D scene representations is essential for effective embodied exploration and reasoning, especially in complex environments over extended periods. Existing representations, such as object-centric 3D scene graphs, oversimplify spatial relationships by modeling scenes as isolated objects with restrictive textual relationships, making it difficult to address queries requiring nuanced spatial understanding. Moreover, these representations lack natural mechanisms for active exploration and memory management, hindering their application to lifelong autonomy. In this work, we propose 3D-Mem, a novel 3D scene memory framework for embodied agents. 3D-Mem employs informative multi-view images, termed Memory Snapshots, to represent the scene and capture rich visual information of explored regions. It further integrates frontier-based exploration by introducing Frontier Snapshots-glimpses of unexplored areas-enabling agents to make informed decisions by considering both known and potential new information. To support lifelong memory in active exploration settings, we present an incremental construction pipeline for 3D-Mem, as well as a memory retrieval technique for memory management. Experimental results on three benchmarks demonstrate that 3D-Mem significantly enhances agents' exploration and reasoning capabilities in 3D environments, highlighting its potential for advancing applications in embodied AI.

As large language models (LLMs) demonstrate increasingly advanced capabilities, aligning their behaviors with human values and preferences becomes crucial for their wide adoption. While previous research focuses on general alignment to principles such as helpfulness, harmlessness, and honesty, the need to account for individual and diverse preferences has been largely overlooked, potentially undermining customized human experiences. To address this gap, we train LLMs that can ''interact to align'', essentially cultivating the meta-skill of LLMs to implicitly infer the unspoken personalized preferences of the current user through multi-turn conversations, and then dynamically align their following behaviors and responses to these inferred preferences. Our approach involves establishing a diverse pool of 3,310 distinct user personas by initially creating seed examples, which are then expanded through iterative self-generation and filtering. Guided by distinct user personas, we leverage multi-LLM collaboration to develop a multi-turn preference dataset containing 3K+ multi-turn conversations in tree structures. Finally, we apply supervised fine-tuning and reinforcement learning to enhance LLMs using this dataset. For evaluation, we establish the ALOE (ALign With CustOmized PrEferences) benchmark, consisting of 100 carefully selected examples and well-designed metrics to measure the customized alignment performance during conversations. Experimental results demonstrate the effectiveness of our method in enabling dynamic, personalized alignment via interaction.

A variety of parameter-efficient fine-tuning methods have emerged in recent years, enabling an increasing number of institutions and researchers to fine-tune LLMs to meet their specific requirements. Adapters (Rebuffi et al., 2017; Houlsby et al., 2019; Lin et al., 2020; Karimi Mahabadi et al., 2021b;a) enable parameter-efficient fine-tuning by inserting trainable layers into LLMs while keeping other layers frozen. However, this approach also introduces additional inference latency. BitFit (Zaken et al., 2021) selectively tunes only the biases within the network, significantly reducing the number of parameters involved in fine-tuning. Prompt tuning achieves parameter-efficient fine-tuning by optimizing a set of new input tokens or prompts for each task (Li & Liang, 2021; Lester et al., 2021; Hambardzumyan et al., 2021; Liu et al., 2023). Hu et al. (2022) introduced LoRA, proposing that weight updates are low-rank and can be expressed as the product of two low-rank matrices. Furthermore, the trainable parameters can be merged with the original weights, eliminating additional inference latency. Recent studies combined parameter-efficient fine-tuning methods with quantization to enhance memory efficiency during the fine-tuning of LLMs (Kwon et al., 2022; Dettmers et al., 2023; Chai et al., 2023; Xu et al., 2023). And DoRA (Liu et al., 2024), or Weight-Decomposed Low-Rank Adaptation, is a parameterefficient fine-tuning method designed to enhance learning capacity and stability by decomposing pre-trained weights into magnitude and direction components, leveraging LoRA for directional updates, and achieving superior performance across tasks without additional inference costs.

This paper proposes a look ahead text understanding problem with look ahead section identification (LASI) as an example. This problem may appear in generative AI as well as human interactions, where we want to understand the direction of a developing text or conversation. We tackle the problem using transformer-based LLMs. We show that LASI is more challenging than classic section identification (SI). We argue that both bidirectional contextual information (e.g., BERT) and unidirectional predictive ability (e.g., GPT) will benefit the task. We propose two approaches to stitch together BERT and GPT. Experiments show that our approach outperforms the established models, especially when there is noise in the text (which is often the case for developing text in generative AI). Our paper sheds light on other look ahead text understanding tasks that are important to social media, such as look ahead sentiment classification, and points out the opportunities to leverage pre-trained LLMs through stitching.

Large language models (LLMs) have significantly advanced natural language processing, excelling in areas like text generation, summarization, and question-answering. Despite their capabilities, these models face challenges when fine-tuned on small, domain-specific datasets, often struggling to generalize and deliver accurate results with unfamiliar inputs. To tackle this issue, we introduce RIRO, a novel two-layer architecture designed to improve performance in data-scarce environments. The first layer leverages advanced prompt engineering to reformulate inputs, ensuring better alignment with training data, while the second layer focuses on refining outputs to minimize inconsistencies. Through fine-tuning models like Phi-2, Falcon 7B, and Falcon 1B, with Phi-2 outperforming the others. Additionally, we introduce a benchmark using evaluation metrics such as cosine similarity, Levenshtein distance, BLEU score, ROUGE-1, ROUGE-2, and ROUGE-L. While these advancements improve performance, challenges like computational demands and overfitting persist, limiting the potential of LLMs in data-scarce, high-stakes environments such as healthcare, legal documentation, and software testing.

Recent advancements in generative AI systems have raised concerns about academic integrity among educators. Beyond excelling at solving programming problems and text-based multiple-choice questions, recent research has also found that large multimodal models (LMMs) can solve Parsons problems based only on an image. However, such problems are still inherently text-based and rely on the capabilities of the models to convert the images of code blocks to their corresponding text. In this paper, we further investigate the capabilities of LMMs to solve graph and tree data structure problems based only on images. To achieve this, we computationally construct and evaluate a novel benchmark dataset comprising 9,072 samples of diverse graph and tree data structure tasks to assess the performance of the GPT-4o, GPT-4v, Gemini 1.5 Pro, Gemini 1.5 Flash, Gemini 1.0 Pro Vision, and Claude 3 model families. GPT-4o and Gemini 1.5 Flash performed best on trees and graphs respectively. GPT-4o achieved 87.6% accuracy on tree samples, while Gemini 1.5 Flash, achieved 56.2% accuracy on graph samples. Our findings highlight the influence of structural and visual variations on model performance. This research not only introduces an LMM benchmark to facilitate replication and further exploration but also underscores the potential of LMMs in solving complex computing problems, with important implications for pedagogy and assessment practices.

As large language models (LLMs) grow in size, traditional full fine-tuning becomes increasingly impractical due to its high computational and storage costs. Although popular parameter-efficient fine-tuning methods, such as LoRA, have significantly reduced the number of tunable parameters, there is still room for further optimization. In this work, we propose ASLoRA, a cross-layer parameter-sharing strategy combining global sharing with partial adaptive sharing. Specifically, we share the low-rank matrix A across all layers and adaptively merge matrix B during training. This sharing mechanism not only mitigates overfitting effectively but also captures inter-layer dependencies, significantly enhancing the model's representational capability. We conduct extensive experiments on various NLP tasks, showing that ASLoRA outperforms LoRA while using less than 25% of the parameters, highlighting its flexibility and superior parameter efficiency. Furthermore, in-depth analyses of the adaptive sharing strategy confirm its significant advantages in enhancing both model flexibility and task adaptability.