We present FIMO, an innovative dataset comprising formal mathematical problem statements sourced from the International Mathematical Olympiad (IMO) Shortlisted Problems. Designed to facilitate advanced automated theorem proving at the IMO level, FIMO is currently tailored for the Lean formal language. It comprises 149 formal problem statements, accompanied by both informal problem descriptions and their corresponding LaTeX-based informal proofs. Through initial experiments involving GPT-4, our findings underscore the existing limitations in current methodologies, indicating a substantial journey ahead before achieving satisfactory IMO-level automated theorem proving outcomes.

While language models are powerful and versatile, they often fail to address highly complex problems. This is because solving complex problems requires deliberate thinking, which has been only minimally guided during training. In this paper, we propose a new method called Cumulative Reasoning (CR), which employs language models in a cumulative and iterative manner to emulate human thought processes. By decomposing tasks into smaller components, CR streamlines the problem-solving process, rendering it both more manageable and effective. For logical inference tasks, CR consistently outperforms existing methods with an improvement up to 9.3%, and achieves an accuracy of 98.04% on the curated FOLIO wiki dataset. In the context of the Game of 24, CR achieves an accuracy of 98%, which signifies a substantial enhancement of 24% over the previous state-of-the-art method. Finally, on the MATH dataset, we establish new state-of-the-art results with 58.0% overall accuracy, surpassing the previous best approach by a margin of 4.2%, and achieving 43% relative improvement on the hardest level 5 problems (22.4% to 32.1%). Additionally, we expand the concept of Cumulative Reasoning to incorporate a Python code environment, deliberately omitting external aids such as retrieval and web browsing and focusing solely on the LLM's intrinsic reasoning capabilities within a Python code environment. Our experiments in this setting yielded impressive results, with an overall accuracy of 72.2% on the MATH dataset, significantly outperforming the PAL method with 38.8% relative improvement. Code is available at https://github.com/iiis-ai/cumulative-reasoning.

Discovering new materials is essential to solve challenges in climate change, sustainability and healthcare. A typical task in materials discovery is to search for a material in a database which maximises the value of a function. That function is often expensive to evaluate, and can rely upon a simulation or an experiment. Here, we introduce SyMDis, a sample efficient optimisation method based on symbolic learning, that discovers near-optimal materials in a large database. SyMDis performs comparably to a state-of-the-art optimiser, whilst learning interpretable rules to aid physical and chemical verification. Furthermore, the rules learned by SyMDis generalise to unseen datasets and return high performing candidates in a zero-shot evaluation, which is difficult to achieve with other approaches.

In real-world tasks, there is usually a large amount of unlabeled data and labeled data. The task of combining the two to learn is known as semi-supervised learning. Experts can use logical rules to label unlabeled data, but this operation is costly. The combination of perception and reasoning has a good effect in processing such semi-supervised tasks with domain knowledge. However, acquiring domain knowledge and the correction, reduction and generation of rules remain complex problems to be solved. Rough set theory is an important method for solving knowledge processing in information systems. In this paper, we propose a rule general abductive learning by rough set (RS-ABL). By transforming the target concept and sub-concepts of rules into information tables, rough set theory is used to solve the acquisition of domain knowledge and the correction, reduction and generation of rules at a lower cost. This framework can also generate more extensive negative rules to enhance the breadth of the knowledge base. Compared with the traditional semi-supervised learning method, RS-ABL has higher accuracy in dealing with semi-supervised tasks.

Propositional model counting (#SAT) can be solved efficiently when the input formula is in deterministic decomposable negation normal form (d-DNNF). Translating an arbitrary formula into a representation that allows inference tasks, such as counting, to be performed efficiently, is called knowledge compilation. Top-down knowledge compilation is a state-of-the-art technique for solving #SAT problems that leverages the traces of exhaustive DPLL search to obtain d-DNNF representations. While knowledge compilation is well studied for propositional approaches, knowledge compilation for the (quantifier free) counting modulo theory setting (#SMT) has been studied to a much lesser degree. In this paper, we discuss compilation strategies for #SMT. We specifically advocate for a top-down compiler based on the traces of exhaustive DPLL(T) search.

Transformer-based language models have been shown to be highly effective for several NLP tasks. In this paper, we consider three transformer models, BERT, RoBERTa, and XLNet, in both small and large versions, and investigate how faithful their representations are with respect to the semantic content of texts. We formalize a notion of semantic faithfulness, in which the semantic content of a text should causally figure in a model's inferences in question answering. We then test this notion by observing a model's behavior on answering questions about a story after performing two novel semantic interventions: deletion intervention and negation intervention. While transformer models achieve high performance on standard question answering tasks, we show that they fail to be semantically faithful once we perform these interventions for a significant number of cases (~50% for deletion intervention, and ~20% drop in accuracy for negation intervention). We then propose an intervention-based training regime that can mitigate the undesirable effects for deletion intervention by a significant margin (from ~ 50% to ~6%). We analyze the inner-workings of the models to better understand the effectiveness of intervention-based training for deletion intervention. But we show that this training does not attenuate other aspects of semantic unfaithfulness such as the models' inability to deal with negation intervention or to capture the predicate-argument structure of texts. We also test InstructGPT, via prompting, for its ability to handle the two interventions and to capture predicate-argument structure. While InstructGPT models do achieve very high performance on predicate-argument structure task, they fail to respond adequately to our deletion and negation interventions.

DisCoCat [1, 2] (Categorical Compositional Distributional) models are structure-preserving maps which send grammatical types to vector spaces and grammatical structures to linear maps. Concretely, the meaning of words is given by tensors with shapes induced by their grammatical types; the meaning of sentences is given by contracting the tensor networks induced by their grammatical structure. String diagrams provide an intuitive graphical language to visualise and reason formally about the evaluation of DisCoCat models; which can be formalised in terms of functors F: G Vect from the category generated by a formal grammar G to the monoidal category Vect of vector spaces and linear maps with the tensor product [3, 2.5]. Although this functorial definition applies equally to any kind of formal grammar, most of the DisCoCat literature focuses on pregroup grammars and more generally on categorial grammars such as the Lambek calculus [4, 5] and combinatory categorial grammars (CCG) [6]. In that case, G is a closed monoidal category and the DisCoCat models F: G Vect map grammatical structures to the closed structure of Vect in a canonical way. In practice, this means that once the meaning of each word is computed from a dataset, the meaning of any new grammatical sentence can be computed automatically from its grammatical structure.

Deep reinforcement learning excels in various domains but lacks generalizability and interoperability. Programmatic RL methods (Trivedi et al., 2021; Liu et al., 2023) reformulate solving RL tasks as synthesizing interpretable programs that can be executed in the environments. Despite encouraging results, these methods are limited to short-horizon tasks. On the other hand, representing RL policies using state machines (Inala et al., 2020) can inductively generalize to long-horizon tasks; however, it struggles to scale up to acquire diverse and complex behaviors. This work proposes Program Machine Policies (POMPs), which bridge the advantages of programmatic RL and state machine policies, allowing for the representation of complex behaviors and the address of long-term tasks. Specifically, we introduce a method that can retrieve a set of effective, diverse, compatible programs. Then, we use these programs as modes of a state machine and learn a transition function to transition among mode programs, allowing for capturing long-horizon repetitive behaviors. Our proposed framework outperforms programmatic RL and deep RL baselines on various tasks and demonstrates the ability to generalize to even longer horizons without any fine-tuning inductively. Ablation studies justify the effectiveness of our proposed search algorithm for retrieving a set of programs as modes.

The control design tools for linear systems typically involves pole placement and computing Lyapunov functions which are useful for ensuring stability. But given higher requirements on control design, a designer is expected to satisfy other specification such as safety or temporal logic specification as well, and a naive control design might not satisfy such specification. A control designer can employ model checking as a tool for checking safety and obtain a counterexample in case of a safety violation. While several scalable techniques for verification have been developed for safety verification of linear dynamical systems, such tools merely act as decision procedures to evaluate system safety and, consequently, yield a counterexample as an evidence to safety violation. However these model checking methods are not geared towards discovering corner cases or re-using verification artifacts for another sub-optimal safety specification. In this paper, we describe a technique for obtaining complete characterization of counterexamples for a safety violation in linear systems. The proposed technique uses the reachable set computed during safety verification for a given temporal logic formula, performs constraint propagation, and represents all modalities of counterexamples using a binary decision diagram (BDD). We introduce an approach to dynamically determine isomorphic nodes for obtaining a considerably reduced (in size) decision diagram. A thorough experimental evaluation on various benchmarks exhibits that the reduction technique achieves up to $67\%$ reduction in the number of nodes and $75\%$ reduction in the width of the decision diagram.

With the increase in industrial applications using Answer Set Programming, the need for formal verification tools, particularly for critical applications, has also increased. During the program optimisation process, it would be desirable to have a tool which can automatically verify whether an optimised subprogram can replace the original subprogram. Formally this corresponds to the problem of verifying the strong equivalence of two programs. In order to do so, the translation tool anthem was developed. It can be used in conjunction with an automated theorem prover for classical logic to verify that two programs are strongly equivalent. With the current version of anthem, only the strong equivalence of positive programs with a restricted input language can be verified. This is a result of the translation $\tau^*$ implemented in anthem that produces formulas in the logic of here-and-there, which coincides with classical logic only for positive programs. This thesis extends anthem in order to overcome these limitations. First, the transformation $\sigma^*$ is presented, which transforms formulas from the logic of here-and-there to classical logic. A theorem formalises how $\sigma^*$ can be used to express equivalence in the logic of here-and-there in classical logic. Second, the translation $\tau^*$ is extended to programs containing pools. Another theorem shows how $\sigma^*$ can be combined with $\tau^*$ to express the strong equivalence of two programs in classical logic. With $\sigma^*$ and the extended $\tau^*$, it is possible to express the strong equivalence of logic programs containing negation, simple choices, and pools. Both the extended $\tau^*$ and $\sigma^*$ are implemented in a new version of anthem. Several examples of logic programs containing pools, negation, and simple choice rules, which the new version of anthem can translate to classical logic, are presented. Some a...