Due to the vast design space of molecules, generating molecules conditioned on a specific sub-structure relevant to a particular function or therapeutic target is a crucial task in computer-aided drug design. Existing works mainly focus on specific tasks, such as linker design or scaffold hopping, each task requires training a model from scratch, and many well-pretrained De Novo molecule generation model parameters are not effectively utilized. To this end, we propose a two-stage training approach, consisting of condition learning and condition optimization. In the condition learning stage, we adopt the idea of ControlNet and design some meaningful adjustments to make the unconditional generative model learn sub-structure conditioned generation. In the condition optimization stage, by using human preference learning, we further enhance the stability and robustness of sub-structure control. In our experiments, only trained on randomly partitioned sub-structure data, the proposed method outperforms previous techniques by generating more valid and diverse molecules. Our method is easy to implement and can be quickly applied to various pre-trained molecule generation models.

Flow-based generative models (FMs) have rapidly advanced as a method for mapping noise to data, its efficient training and sampling process makes it widely applicable in various fields. FMs can be viewed as a variant of diffusion models (DMs). At the same time, previous studies have shown that DMs are vulnerable to Trojan/Backdoor attacks, a type of output manipulation attack triggered by a maliciously embedded pattern at model input. We found that Trojan attacks on generative models are essentially equivalent to image transfer tasks from the backdoor distribution to the target distribution, the unique ability of FMs to fit any two arbitrary distributions significantly simplifies the training and sampling setups for attacking FMs, making them inherently natural targets for backdoor attacks. In this paper, we propose TrojFlow, exploring the vulnerabilities of FMs through Trojan attacks. In particular, we consider various attack settings and their combinations and thoroughly explore whether existing defense methods for DMs can effectively defend against our proposed attack scenarios. We evaluate TrojFlow on CIFAR-10 and CelebA datasets, our experiments show that our method can compromise FMs with high utility and specificity, and can easily break through existing defense mechanisms.

Ruiyi Wang, Haofei Yu, Wenxin Zhang, Zhengyang Qi, Maarten Sap, Yonatan Bisk, Graham Neubig, Hao Zhu. Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers). 2024.

Advances in multimodal models have greatly improved how interactions relevant to various tasks are modeled.Today's multimodal models mainly focus on the correspondence between images and text, using this for tasks like imagetext matching.However, this covers only a subset of real-world interactions.Novel interactions, such as sarcasm expressed through opposing spoken words and gestures or humor expressed through utterances and tone of voice, remain challenging.In this paper, we introduce an approach to enhance multimodal models, which we call Multimodal Mixtures of Experts (MMOE).The key idea in MMOE is to train separate expert models for each type of multimodal interaction, such as redundancy present in both modalities, uniqueness in one modality, or synergy that emerges when both modalities are fused.On a sarcasm detection task (MUStARD) and a humor detection task (URFunny), we obtain new state-of-the-art results.MMOE is also able to be applied to various types of models to gain improvement.

*Humans are social beings*; we pursue social goals in our daily interactions, which is a crucial aspect of social intelligence. Yet, AI systems' abilities in this realm remain elusive. We present SOTOPIA, an open-ended environment to simulate complex social interactions between artificial agents and evaluate their social intelligence. In our environment, agents role-play and *interact* under a wide variety of scenarios; they coordinate, collaborate, exchange, and compete with each other to achieve complex social goals. We simulate the role-play interaction between LLM-based agents and humans within this task space and evaluate their performance with a holistic evaluation framework called SOTOPIA-Eval. With SOTOPIA, we find significant differences between these models in terms of their social intelligence, and we identify a subset of SOTOPIA scenarios, SOTOPIA-hard, that is generally challenging for all models. We find that on this subset, GPT-4 achieves a significantly lower goal completion rate than humans and struggles to exhibit social commonsense reasoning and strategic communication skills. These findings demonstrate SOTOPIA's promise as a general platform for research on evaluating and improving social intelligence in artificial agents.

Social media (i.e., Reddit) users are overloaded with people’s opinions when viewing discourses about divisive topics. Traditional user interfaces in such media present those opinions in a linear structure, which can limit users in viewing diverse social opinions at scale. Prior work has recognized this limitation, that the linear structure can reinforce biases, where a certain point of view becomes widespread simply because many viewers seem to believe it. This limitation can make it difficult for users to have a truly conversational mode of mediated discussion. Thus, when designing a user interface for viewing people’s opinions, we should consider ways to mitigate selective exposure to information and polarization of opinions. We conducted a needs-finding study with 11 Reddit users, who follow climate change threads and make posts and comments regularly. In the study, we aimed to understand key limitations in people viewing online controversial discourses and to extract design implications to address these problems. Our findings discuss potential future directions to address these problems.