Few-Shot Multilingual Open-Domain QA from 5 Examples

Recent approaches to multilingual open-domain question answering (MLODQA) have achieved promising results given abundant language-specific training data. However, the considerable annotation cost limits the application of these methods for underrepresented languages. We introduce a \emph{few-shot learning} approach to synthesise large-scale multilingual data from large language models (LLMs). Our method begins with large-scale self-supervised pre-training using WikiData, followed by training on high-quality synthetic multilingual data generated by prompting LLMs with few-shot supervision. The final model, \textsc{FsModQA}, significantly outperforms existing few-shot and supervised baselines in MLODQA and cross-lingual and monolingual retrieval. We further show our method can be extended for effective zero-shot adaptation to new languages through a \emph{cross-lingual prompting} strategy with only English-supervised data, making it a general and applicable solution for MLODQA tasks without costly large-scale annotation.

FantasyID: Face Knowledge Enhanced ID-Preserving Video Generation

Tuning-free approaches adapting large-scale pre-trained video diffusion models for identity-preserving text-to-video generation (IPT2V) have gained popularity recently due to their efficacy and scalability. However, significant challenges remain to achieve satisfied facial dynamics while keeping the identity unchanged. In this work, we present a novel tuning-free IPT2V framework by enhancing face knowledge of the pre-trained video model built on diffusion transformers (DiT), dubbed FantasyID. Essentially, 3D facial geometry prior is incorporated to ensure plausible facial structures during video synthesis. To prevent the model from learning copy-paste shortcuts that simply replicate reference face across frames, a multi-view face augmentation strategy is devised to capture diverse 2D facial appearance features, hence increasing the dynamics over the facial expressions and head poses. Additionally, after blending the 2D and 3D features as guidance, instead of naively employing cross-attention to inject guidance cues into DiT layers, a learnable layer-aware adaptive mechanism is employed to selectively inject the fused features into each individual DiT layers, facilitating balanced modeling of identity preservation and motion dynamics. Experimental results validate our model's superiority over the current tuning-free IPT2V methods.

Intrinsic Cramér-Rao Bound based 6D Localization and Tracking for 5G/6G Systems

Localization and tracking are critical components of integrated sensing and communication (ISAC) systems, enhancing resource management, beamforming accuracy, and overall system reliability through precise sensing. Due to the high path loss of the high-frequency systems, antenna arrays are required at the transmitter and receiver sides for beamforming gain. However, beam misalignment may occur, which requires accurate tracking of the six-dimensional (6D) state, namely, 3D position and 3D orientation. In this work, we first address the challenge that the rotation matrix, being part of the Lie group rather than Euclidean space, necessitates the derivation of the ICRB for an intrinsic performance benchmark. Then, leveraging the derived ICRB, we develop two filters-one utilizing pose fusion and the other employing error-state Kalman filter to estimate the UE's 6D state for different computational resource consumption and accuracy requirements. Simulation results validate the ICRB and assess the performance of the proposed filters, demonstrating their effectiveness and improved accuracy in 6D state tracking.

Franken-Adapter: Cross-Lingual Adaptation of LLMs by Embedding Surgery

The capabilities of Large Language Models (LLMs) in low-resource languages lag far behind those in English, making their universal accessibility a significant challenge. To alleviate this, we present $\textit{Franken-Adapter}$, a modular language adaptation approach for decoder-only LLMs with embedding surgery. Our method begins by creating customized vocabularies for target languages and performing language adaptation through embedding tuning on multilingual data. These pre-trained embeddings are subsequently integrated with LLMs that have been instruction-tuned on English alignment data to enable zero-shot cross-lingual transfer. Our experiments on $\texttt{Gemma2}$ models with up to 27B parameters demonstrate improvements of up to 20% across 96 languages, spanning both discriminative and generative tasks, with minimal regressions ($<$1%) in English. Further in-depth analysis reveals the critical role of customizing tokenizers in enhancing language adaptation, while boosting inference efficiency. Additionally, we show the versatility of our method by achieving a 14% improvement over a math-optimized LLM across 20 languages, offering a modular solution to transfer reasoning abilities across languages post hoc.

Human-In-the-Loop Software Development Agents

Recently, Large Language Models (LLMs)-based multi-agent paradigms for software engineering are introduced to automatically resolve software development tasks (e.g., from a given issue to source code). However, existing work is evaluated based on historical benchmark datasets, does not consider human feedback at each stage of the automated software development process, and has not been deployed in practice. In this paper, we introduce a Human-in-the-loop LLM-based Agents framework (HULA) for software development that allows software engineers to refine and guide LLMs when generating coding plans and source code for a given task. We design, implement, and deploy the HULA framework into Atlassian JIRA for internal uses. Through a multi-stage evaluation of the HULA framework, Atlassian software engineers perceive that HULA can minimize the overall development time and effort, especially in initiating a coding plan and writing code for straightforward tasks. On the other hand, challenges around code quality are raised to be solved in some cases. We draw lessons learned and discuss opportunities for future work, which will pave the way for the advancement of LLM-based agents in software development.

Hunyuan-Large: An Open-Source MoE Model with 52 Billion Activated Parameters by Tencent

In this paper, we introduce Hunyuan-Large, which is currently the largest open-source Transformer-based mixture of experts model, with a total of 389 billion parameters and 52 billion activation parameters, capable of handling up to 256K tokens. We conduct a thorough evaluation of Hunyuan-Large's superior performance across various benchmarks including language understanding and generation, logical reasoning, mathematical problem-solving, coding, long-context, and aggregated tasks, where it outperforms LLama3.1-70B and exhibits comparable performance when compared to the significantly larger LLama3.1-405B model. Key practice of Hunyuan-Large include large-scale synthetic data that is orders larger than in previous literature, a mixed expert routing strategy, a key-value cache compression technique, and an expert-specific learning rate strategy. Additionally, we also investigate the scaling laws and learning rate schedule of mixture of experts models, providing valuable insights and guidances for future model development and optimization. The code and checkpoints of Hunyuan-Large are released to facilitate future innovations and applications. Codes: https://github.com/Tencent/Hunyuan-Large Models: https://huggingface.co/tencent/Tencent-Hunyuan-Large

DiP-GO: A Diffusion Pruner via Few-step Gradient Optimization

Diffusion models have achieved remarkable progress in the field of image generation due to their outstanding capabilities. However, these models require substantial computing resources because of the multi-step denoising process during inference. While traditional pruning methods have been employed to optimize these models, the retraining process necessitates large-scale training datasets and extensive computational costs to maintain generalization ability, making it neither convenient nor efficient. Recent studies attempt to utilize the similarity of features across adjacent denoising stages to reduce computational costs through simple and static strategies. However, these strategies cannot fully harness the potential of the similar feature patterns across adjacent timesteps. In this work, we propose a novel pruning method that derives an efficient diffusion model via a more intelligent and differentiable pruner. At the core of our approach is casting the model pruning process into a SubNet search process. Specifically, we first introduce a SuperNet based on standard diffusion via adding some backup connections built upon the similar features. We then construct a plugin pruner network and design optimization losses to identify redundant computation. Finally, our method can identify an optimal SubNet through few-step gradient optimization and a simple post-processing procedure. We conduct extensive experiments on various diffusion models including Stable Diffusion series and DiTs. Our DiP-GO approach achieves 4.4 x speedup for SD-1.5 without any loss of accuracy, significantly outperforming the previous state-of-the-art methods.

Channel Estimation, Interpolation and Extrapolation in Doubly-dispersive Channels

The OTFS (Orthogonal Time Frequency Space) is widely acknowledged for its ability to combat Doppler spread in time-varying channels. In this paper, another advantage of OTFS over OFDM (Orthogonal Frequency Division Multiplexing) will be demonstrated: much reduced channel training overhead. Specifically, the sparsity of the channel in delay-Doppler (D-D) domain implies strong correlation of channel gains in time-frequency (T-F) domain, which can be harnessed to reduce channel training overhead through interpolation. An immediate question is how much training overhead is needed in doubly-dispersive channels? A conventional belief is that the overhead is only dependent on the product of delay and Doppler spreads, but we will show that it's also dependent on the T-F window size. The finite T-F window leads to infinite spreading in D-D domain, and aliasing will be inevitable after sampling in T-F domain. Two direct consequences of the aliasing are increased channel training overhead and interference. Another factor contributing to channel estimation error is the inter-symbol-carrier-interference (ISCI), resulting from the uncertainty principle. Both aliasing and ISCI are considered in channel modelling, a low-complexity algorithm is proposed for channel estimation and interpolation through FFT. A large T-F window is necessary for reduced channel training overhead and aliasing, but increases processing delay. Fortunately, we show that the proposed algorithm can be implemented in a pipeline fashion. Further more, we showed that data-aided channel tracking is possible in D-D domain to further reduce the channel estimation frequency, i.e., channel extrapolation. The impacts of aliasing and ISCI on channel interpolation error are analyzed.

Controllable Longer Image Animation with Diffusion Models

Generating realistic animated videos from static images is an important area of research in computer vision. Methods based on physical simulation and motion prediction have achieved notable advances, but they are often limited to specific object textures and motion trajectories, failing to exhibit highly complex environments and physical dynamics. In this paper, we introduce an open-domain controllable image animation method using motion priors with video diffusion models. Our method achieves precise control over the direction and speed of motion in the movable region by extracting the motion field information from videos and learning moving trajectories and strengths. Current pretrained video generation models are typically limited to producing very short videos, typically less than 30 frames. In contrast, we propose an efficient long-duration video generation method based on noise reschedule specifically tailored for image animation tasks, facilitating the creation of videos over 100 frames in length while maintaining consistency in content scenery and motion coordination. Specifically, we decompose the denoise process into two distinct phases: the shaping of scene contours and the refining of motion details. Then we reschedule the noise to control the generated frame sequences maintaining long-distance noise correlation. We conducted extensive experiments with 10 baselines, encompassing both commercial tools and academic methodologies, which demonstrate the superiority of our method. Our project page: https://wangqiang9.github.io/Controllable.github.io/

Integrated Communications and Localization for Massive MIMO LEO Satellite Systems

Integrated communications and localization (ICAL) will play an important part in future sixth generation (6G) networks for the realization of Internet of Everything (IoE) to support both global communications and seamless localization. Massive multiple-input multiple-output (MIMO) low earth orbit (LEO) satellite systems have great potential in providing wide coverage with enhanced gains, and thus are strong candidates for realizing ubiquitous ICAL. In this paper, we develop a wideband massive MIMO LEO satellite system to simultaneously support wireless communications and localization operations in the downlink. In particular, we first characterize the signal propagation properties and derive a localization performance bound. Based on these analyses, we focus on the hybrid analog/digital precoding design to achieve high communication capability and localization precision. Numerical results demonstrate that the proposed ICAL scheme supports both the wireless communication and localization operations for typical system setups.