In this paper, we propose a novel framework for enhancing visual comprehension in autonomous driving systems by integrating visual language models (VLMs) with additional visual perception module specialised in object detection. We extend the Llama-Adapter architecture by incorporating a YOLOS-based detection network alongside the CLIP perception network, addressing limitations in object detection and localisation. Our approach introduces camera ID-separators to improve multi-view processing, crucial for comprehensive environmental awareness. Experiments on the DriveLM visual question answering challenge demonstrate significant improvements over baseline models, with enhanced performance in ChatGPT scores, BLEU scores, and CIDEr metrics, indicating closeness of model answer to ground truth. Our method represents a promising step towards more capable and interpretable autonomous driving systems. Possible safety enhancement enabled by detection modality is also discussed.

The remarkable performance achieved by Large Language Models (LLM) has driven research efforts to leverage them for a wide range of tasks and input modalities. In speech-to-text (S2T) tasks, the emerging solution consists of projecting the output of the encoder of a Speech Foundational Model (SFM) into the LLM embedding space through an adapter module. However, no work has yet investigated how much the downstream-task performance depends on each component (SFM, adapter, LLM) nor whether the best design of the adapter depends on the chosen SFM and LLM. To fill this gap, we evaluate the combination of 5 adapter modules, 2 LLMs (Mistral and Llama), and 2 SFMs (Whisper and SeamlessM4T) on two widespread S2T tasks, namely Automatic Speech Recognition and Speech Translation. Our results demonstrate that the SFM plays a pivotal role in downstream performance, while the adapter choice has moderate impact and depends on the SFM and LLM.

Zyda-2 was used to train our Zamba2 series of models which are state-of-the-art for their weight class. We build Zyda-2 by collating high-quality open-source tokens such as FineWeb and DCLM, then distilling them to the highest-quality subset via cross-deduplication and model-based quality filtering. Zyda-2 is released under a permissive open license, and is available at https://huggingface.co/datasets/Zyphra/Zyda-2.

Machine learning methods have a groundbreaking impact in many application domains, but their application on real robotic platforms is still limited. Despite the many challenges associated with combining machine learning technology with robotics, robot learning remains one of the most promising directions for enhancing the capabilities of robots. When deploying learning-based approaches on real robots, extra effort is required to address the challenges posed by various real-world factors. To investigate the key factors influencing real-world deployment and to encourage original solutions from different researchers, we organized the Robot Air Hockey Challenge at the NeurIPS 2023 conference. We selected the air hockey task as a benchmark, encompassing low-level robotics problems and high-level tactics. Different from other machine learning-centric benchmarks, participants need to tackle practical challenges in robotics, such as the sim-to-real gap, low-level control issues, safety problems, real-time requirements, and the limited availability of real-world data. Furthermore, we focus on a dynamic environment, removing the typical assumption of quasi-static motions of other real-world benchmarks. The competition's results show that solutions combining learning-based approaches with prior knowledge outperform those relying solely on data when real-world deployment is challenging. Our ablation study reveals which real-world factors may be overlooked when building a learning-based solution. The successful real-world air hockey deployment of best-performing agents sets the foundation for future competitions and follow-up research directions.

Sandboxes and other dynamic analysis processes are prevalent in malware detection systems nowadays to enhance the capability of detecting 0-day malware. Therefore, techniques of anti-dynamic analysis (TADA) are prevalent in modern malware samples, and sandboxes can suffer from false negatives and analysis failures when analyzing the samples with TADAs. In such cases, human reverse engineers will get involved in conducting dynamic analysis manually (i.e., debugging, patching), which in turn also gets obstructed by TADAs. In this work, we propose a Large Language Model (LLM) based workflow that can pinpoint the location of the TADA implementation in the code, to help reverse engineers place breakpoints used in debugging. Our evaluation shows that we successfully identified the locations of 87.80% known TADA implementations adopted from public repositories. In addition, we successfully pinpoint the locations of TADAs in 4 well-known malware samples that are documented in online malware analysis blogs.

Large Language Models (LLMs) have demonstrated remarkable capabilities in conversational tasks. Embodying an LLM as a virtual human allows users to engage in face-to-face social interactions in Virtual Reality. However, the influence of person- and task-related factors in social interactions with LLM-controlled agents remains unclear. In this study, forty-six participants interacted with a virtual agent whose persona was manipulated as extravert or introvert in three different conversational tasks (small talk, knowledge test, convincing). Social-evaluation, emotional experience, and realism were assessed using ratings. Interactive engagement was measured by quantifying participants' words and conversational turns. Finally, we measured participants' willingness to ask the agent for help during the knowledge test. Our findings show that the extraverted agent was more positively evaluated, elicited a more pleasant experience and greater engagement, and was assessed as more realistic compared to the introverted agent. Whereas persona did not affect the tendency to ask for help, participants were generally more confident in the answer when they had help of the LLM. Variation of personality traits of LLM-controlled embodied virtual agents, therefore, affects social-emotional processing and behavior in virtual interactions. Embodied virtual agents allow the presentation of naturalistic social encounters in a virtual environment.

Large language models (LLMs) represent a promising, but controversial, tool in aiding scientific peer review. This study evaluates the usefulness of LLMs in a conference setting as a tool for vetting paper submissions against submission standards. We conduct an experiment at the 2024 Neural Information Processing Systems (NeurIPS) conference, where 234 papers were voluntarily submitted to an "LLM-based Checklist Assistant." This assistant validates whether papers adhere to the author checklist used by NeurIPS, which includes questions to ensure compliance with research and manuscript preparation standards. Evaluation of the assistant by NeurIPS paper authors suggests that the LLM-based assistant was generally helpful in verifying checklist completion. In post-usage surveys, over 70% of authors found the assistant useful, and 70% indicate that they would revise their papers or checklist responses based on its feedback. While causal attribution to the assistant is not definitive, qualitative evidence suggests that the LLM contributed to improving some submissions. Survey responses and analysis of re-submissions indicate that authors made substantive revisions to their submissions in response to specific feedback from the LLM. The experiment also highlights common issues with LLMs: inaccuracy (20/52) and excessive strictness (14/52) were the most frequent issues flagged by authors. We also conduct experiments to understand potential gaming of the system, which reveal that the assistant could be manipulated to enhance scores through fabricated justifications, highlighting potential vulnerabilities of automated review tools.

Have you ever tried to use an AI tool like ChatGPT to do some math and found it doesn't always add up? It turns out there's a reason for that. As large language models (LLMs) like OpenAI's ChatGPT become more ubiquitous, people increasingly rely on them for work and research assistance. Yuntian Deng, assistant professor at the David R. Cheriton School of Computer Science, discusses some of the challenges in LLMs' reasoning capabilities, particularly in math, and explores the implications of using these models to aid problem-solving. As I explained in a recent post on X, the latest reasoning variant of ChatGPT o1, struggles with large-digit multiplication, especially when multiplying numbers beyond nine digits. This is a notable improvement over the previous ChatGPT-4o model, which struggled even with four-digit multiplication, but it's still a major flaw.

Large Language Models (LLMs), such as ChatGPT, Phi3 and Llama-3, are leading a significant leap in AI, as they can generalize knowledge from their training to new tasks without fine-tuning. However, their application in the financial domain remains relatively limited. The financial field is inherently complex, requiring a deep understanding across various perspectives, from macro, micro economic trend to quantitative analysis. Motivated by this complexity, a mixture of expert LLMs tailored to specific financial domains could offer a more comprehensive understanding for intricate financial tasks. In this paper, we present the FinTeamExperts, a role-specialized LLM framework structured as a Mixture of Experts (MOEs) for financial analysis. The framework simulates a collaborative team setting by training each model to specialize in distinct roles: Macro Analysts, Micro analysts, and Quantitative Analysts. This role-specific specialization enhances the model's ability to integrate their domain-specific expertise. We achieve this by training three 8-billion parameter models on different corpus, each dedicated to excelling in specific finance-related roles. We then instruct-tune FinTeamExperts on downstream tasks to align with practical financial tasks. The experimental results show that FinTeamExperts outperform all models of the same size and larger on three out of four datasets. On the fourth dataset, which presents a more complex task, FinTeamExperts still surpass all models of the same size. This highlights the success of our role-based specialization approach and the continued training approach for FinTeamExperts.

This paper presents a novel framework for watermarking language models through prompts generated by language models. The proposed approach utilizes a multi-model setup, incorporating a Prompting language model to generate watermarking instructions, a Marking language model to embed watermarks within generated content, and a Detecting language model to verify the presence of these watermarks. Experiments are conducted using ChatGPT and Mistral as the Prompting and Marking language models, with detection accuracy evaluated using a pretrained classifier model. Results demonstrate that the proposed framework achieves high classification accuracy across various configurations, with 95% accuracy for ChatGPT, 88.79% for Mistral. These findings validate the and adaptability of the proposed watermarking strategy across different language model architectures. Hence the proposed framework holds promise for applications in content attribution, copyright protection, and model authentication.