The SIFo Benchmark: Investigating the Sequential Instruction Following Ability of Large Language Models

Following multiple instructions is a crucial ability for large language models (LLMs). Evaluating this ability comes with significant challenges: (i) limited coherence between multiple instructions, (ii) positional bias where the order of instructions affects model performance, and (iii) a lack of objectively verifiable tasks. To address these issues, we introduce a benchmark designed to evaluate models' abilities to follow multiple instructions through sequential instruction following (SIFo) tasks. In SIFo, the successful completion of multiple instructions is verifiable by examining only the final instruction. Our benchmark evaluates instruction following using four tasks (text modification, question answering, mathematics, and security rule following), each assessing different aspects of sequential instruction following. Our evaluation of popular LLMs, both closed-source and open-source, shows that more recent and larger models significantly outperform their older and smaller counterparts on the SIFo tasks, validating the benchmark's effectiveness. All models struggle with following sequences of instructions, hinting at an important lack of robustness of today's language models.

Intelligent Text-Conditioned Music Generation

CLIP (Contrastive Language-Image Pre-Training) is a multimodal neural network trained on (text, image) pairs to predict the most relevant text caption given an image. It has been used extensively in image generation by connecting its output with a generative model such as VQGAN, with the most notable example being OpenAI's DALLE-2. In this project, we apply a similar approach to bridge the gap between natural language and music. Our model is split into two steps: first, we train a CLIP-like model on pairs of text and music over contrastive loss to align a piece of music with its most probable text caption. Then, we combine the alignment model with a music decoder to generate music. To the best of our knowledge, this is the first attempt at text-conditioned deep music generation. Our experiments show that it is possible to train the text-music alignment model using contrastive loss and train a decoder to generate music from text prompts.

Making Recommender Systems More Knowledgeable: A Framework to Incorporate Side Information

Session-based recommender systems typically focus on using only the triplet (user_id, timestamp, item_id) to make predictions of users' next actions. In this paper, we aim to utilize side information to help recommender systems catch patterns and signals otherwise undetectable. Specifically, we propose a general framework for incorporating item-specific side information into the recommender system to enhance its performance without much modification on the original model architecture. Experimental results on several models and datasets prove that with side information, our recommender system outperforms state-of-the-art models by a considerable margin and converges much faster. Additionally, we propose a new type of loss to regularize the attention mechanism used by recommender systems and evaluate its influence on model performance. Furthermore, through analysis, we put forward a few insights on potential further improvements.

Preference Learning Algorithms Do Not Learn Preference Rankings

Preference learning algorithms (e.g., RLHF and DPO) are frequently used to steer LLMs to produce generations that are more preferred by humans, but our understanding of their inner workings is still limited. In this work, we study the conventional wisdom that preference learning trains models to assign higher likelihoods to more preferred outputs than less preferred outputs, measured via $\textit{ranking accuracy}$. Surprisingly, we find that most state-of-the-art preference-tuned models achieve a ranking accuracy of less than 60% on common preference datasets. We furthermore derive the $\textit{idealized ranking accuracy}$ that a preference-tuned LLM would achieve if it optimized the DPO or RLHF objective perfectly. We demonstrate that existing models exhibit a significant $\textit{alignment gap}$ -- $\textit{i.e.}$, a gap between the observed and idealized ranking accuracies. We attribute this discrepancy to the DPO objective, which is empirically and theoretically ill-suited to fix even mild ranking errors in the reference model, and derive a simple and efficient formula for quantifying the difficulty of learning a given preference datapoint. Finally, we demonstrate that ranking accuracy strongly correlates with the empirically popular win rate metric when the model is close to the reference model used in the objective, shedding further light on the differences between on-policy (e.g., RLHF) and off-policy (e.g., DPO) preference learning algorithms.

Dual-Adapter: Training-free Dual Adaptation for Few-shot Out-of-Distribution Detection

We study the problem of few-shot out-of-distribution (OOD) detection, which aims to detect OOD samples from unseen categories during inference time with only a few labeled in-domain (ID) samples. Existing methods mainly focus on training task-aware prompts for OOD detection. However, training on few-shot data may cause severe overfitting and textual prompts alone may not be enough for effective detection. To tackle these problems, we propose a prior-based Training-free Dual Adaptation method (Dual-Adapter) to detect OOD samples from both textual and visual perspectives. Specifically, Dual-Adapter first extracts the most significant channels as positive features and designates the remaining less relevant channels as negative features. Then, it constructs both a positive adapter and a negative adapter from a dual perspective, thereby better leveraging previously outlooked or interfering features in the training dataset. In this way, Dual-Adapter can inherit the advantages of CLIP not having to train, but also excels in distinguishing between ID and OOD samples. Extensive experimental results on four benchmark datasets demonstrate the superiority of Dual-Adapter.

Dual-Robust Integrated Sensing and Communication: Beamforming under CSI Imperfection and Location Uncertainty

A dual-robust design of beamforming is investigated in an integrated sensing and communication (ISAC) system.Existing research on robust ISAC waveform design, while proposing solutions to imperfect channel state information (CSI), generally depends on prior knowledge of the target's approximate location to design waveforms. This approach, however, limits the precision in sensing the target's exact location. In this paper, considering both CSI imperfection and target location uncertainty, a novel framework of joint robust optimization is proposed by maximizing the weighted sum of worst-case data rate and beampattern gain. To address this challenging problem, we propose an efficient two-layer iteration algorithm based on S-Procedure and convex hull. Finally, numerical results verify the effectiveness and performance improvement of our dual-robust algorithm, as well as the trade-off between communication and sensing performance.

APACE: Agile and Perception-Aware Trajectory Generation for Quadrotor Flights

Various perception-aware planning approaches have attempted to enhance the state estimation accuracy during maneuvers, while the feature matchability among frames, a crucial factor influencing estimation accuracy, has often been overlooked. In this paper, we present APACE, an Agile and Perception-Aware trajeCtory gEneration framework for quadrotors aggressive flight, that takes into account feature matchability during trajectory planning. We seek to generate a perception-aware trajectory that reduces the error of visual-based estimator while satisfying the constraints on smoothness, safety, agility and the quadrotor dynamics. The perception objective is achieved by maximizing the number of covisible features while ensuring small enough parallax angles. Additionally, we propose a differentiable and accurate visibility model that allows decomposition of the trajectory planning problem for efficient optimization resolution. Through validations conducted in both a photorealistic simulator and real-world experiments, we demonstrate that the trajectories generated by our method significantly improve state estimation accuracy, with root mean square error (RMSE) reduced by up to an order of magnitude. The source code will be released to benefit the community.

Efficient Online Learning for Networks of Two-Compartment Spiking Neurons

The brain-inspired Spiking Neural Networks (SNNs) have garnered considerable research interest due to their superior performance and energy efficiency in processing temporal signals. Recently, a novel multi-compartment spiking neuron model, namely the Two-Compartment LIF (TC-LIF) model, has been proposed and exhibited a remarkable capacity for sequential modelling. However, training the TC-LIF model presents challenges stemming from the large memory consumption and the issue of gradient vanishing associated with the Backpropagation Through Time (BPTT) algorithm. To address these challenges, online learning methodologies emerge as a promising solution. Yet, to date, the application of online learning methods in SNNs has been predominantly confined to simplified Leaky Integrate-and-Fire (LIF) neuron models. In this paper, we present a novel online learning method specifically tailored for networks of TC-LIF neurons. Additionally, we propose a refined TC-LIF neuron model called Adaptive TC-LIF, which is carefully designed to enhance temporal information integration in online learning scenarios. Extensive experiments, conducted on various sequential benchmarks, demonstrate that our approach successfully preserves the superior sequential modeling capabilities of the TC-LIF neuron while incorporating the training efficiency and hardware friendliness of online learning. As a result, it offers a multitude of opportunities to leverage neuromorphic solutions for processing temporal signals.

Knowledge Graph Assisted Automatic Sports News Writing

In this paper, we present a novel method for automatically generating sports news, which employs a unique algorithm that extracts pivotal moments from live text broadcasts and uses them to create an initial draft of the news. This draft is further refined by incorporating key details and background information from a specially designed sports knowledge graph. This graph contains 5,893 entities, which are classified into three distinct conceptual categories, interconnected through four relationship types, and characterized by 27 unique attributes. In addition, we create a multi-stage learning model by combining convolutional neural networks and a transformer encoder. This model expresses entity-task interactions using convolutional neural networks and enriches entity representations in the query set with the transformer encoder. It also includes a processor to compute matching scores for incomplete triples, addressing few-shot knowledge graph completion problem. The efficiency of this approach has been confirmed through both subjective and objective evaluations of 50 selected test cases, demonstrating its capability in revolutionizing the creation of sports news.

Adaptive Regret for Bandits Made Possible: Two Queries Suffice

Fast changing states or volatile environments pose a significant challenge to online optimization, which needs to perform rapid adaptation under limited observation. In this paper, we give query and regret optimal bandit algorithms under the strict notion of strongly adaptive regret, which measures the maximum regret over any contiguous interval $I$. Due to its worst-case nature, there is an almost-linear $\Omega(|I|^{1-\epsilon})$ regret lower bound, when only one query per round is allowed [Daniely el al, ICML 2015]. Surprisingly, with just two queries per round, we give Strongly Adaptive Bandit Learner (StABL) that achieves $\tilde{O}(\sqrt{n|I|})$ adaptive regret for multi-armed bandits with $n$ arms. The bound is tight and cannot be improved in general. Our algorithm leverages a multiplicative update scheme of varying stepsizes and a carefully chosen observation distribution to control the variance. Furthermore, we extend our results and provide optimal algorithms in the bandit convex optimization setting. Finally, we empirically demonstrate the superior performance of our algorithms under volatile environments and for downstream tasks, such as algorithm selection for hyperparameter optimization.