相关论文: Key-Conditioned Orthonormal Transform Gating (K-OT…

Logic locking has emerged to prevent piracy and overproduction of integrated circuits ever since the split of the design house and manufacturing foundry was established. While there has been a lot of research using a single global key to…

Embodied control increasingly requires models to follow compositional language instructions while reasoning over dynamic visual states. However, current vision-language-action policies and world-action models often couple linguistic…

Activation steering controls LLM behaviour towards target behaviour by intervening in internal representations, yet it often degrades reasoning and retrieval performance. We argue that a primary cause of this trade-off is attention…

Language models (LMs) have been shown to behave unexpectedly post-deployment. For example, new jailbreaks continually arise, allowing model misuse, despite extensive red-teaming and adversarial training from developers. Given most model…

Existing invasive (backdoor) fingerprints suffer from high-perplexity triggers that are easily filtered, fixed response patterns exposed by heuristic detectors, and spurious activations on benign inputs. We introduce \textsc{ForgetMark}, a…

Benefiting from massive corpora and advanced hardware, large language models (LLMs) exhibit remarkable capabilities in language understanding and generation. However, their performance degrades in scenarios where multiple tasks are…

In a zero-trust fabless paradigm, designers are increasingly concerned about hardware-based attacks on the semiconductor supply chain. Logic locking is a design-for-trust method that adds extra key-controlled gates in the circuits to…

Test-Time Adaptation (TTA) enables pre-trained models to adjust to distribution shift by learning from unlabeled test-time streams. However, existing methods typically treat these streams as independent samples, overlooking the supervisory…

Graph transformers achieve strong results on molecular and long-range reasoning tasks, yet remain hampered by over-smoothing (the progressive collapse of node representations with depth) and attention entropy degeneration. We observe that…

This paper aims to efficiently enable Large Language Models (LLMs) to use multimodal tools. Advanced proprietary LLMs, such as ChatGPT and GPT-4, have shown great potential for tool usage through sophisticated prompt engineering.…

Despite the success of Large Language Models (LLMs) on various tasks following human instructions, controlling model generation at inference time poses a persistent challenge. In this paper, we introduce Ctrl-G, an adaptable framework that…

Reorientation (turning in plane) plays a critical role for all robots in any field application, especially those that in confined spaces. While important, reorientation remains a relatively unstudied problem for robots, including limbless…

This paper presents an ontology-aware pretrained language model (OPAL) for end-to-end task-oriented dialogue (TOD). Unlike chit-chat dialogue models, task-oriented dialogue models fulfill at least two task-specific modules: dialogue state…

Adapting large pre-trained models to unseen tasks under tight data and compute budgets remains challenging. Meta-learning approaches explicitly learn good initializations, but they require an additional meta-training phase over many tasks,…

Weight tying is widely used in compact language models to reduce parameters by sharing the token table between the input embedding and the output projection. However, parameter sharing alone does not guarantee a stable token interface:…

In this paper, we propose Ahead-of-Time (AoT) P-Tuning, a novel parameter-efficient fine-tuning method for pre-trained Language Models (LMs) that adds input-dependent bias before each Transformer layer. We evaluate AoT P-Tuning on GLUE and…

The Transformer architecture has been successful across many domains, including natural language processing, computer vision and speech recognition. In keyword spotting, self-attention has primarily been used on top of convolutional or…

Recent studies show that training deep neural networks (DNNs) with Lipschitz constraints are able to enhance adversarial robustness and other model properties such as stability. In this paper, we propose a layer-wise orthogonal training…

Multimodal pretrained models are vulnerable to backdoor attacks, yet most existing methods rely on visual or multimodal triggers, which are impractical since visually embedded triggers rarely occur in real-world data. To overcome this…

Recently, adapting pre-trained models to downstream tasks has attracted increasing interest. Previous Parameter-Efficient-Tuning (PET) methods regard the pre-trained model as an opaque Black Box model, relying purely on data-driven…