@inproceedings{18086, abstract = {Abstract. Continuous group key agreement (CGKA) allows a group of users to maintain a continuously updated shared key in an asynchronous setting where parties only come online sporadically and their messages are relayed by an untrusted server. CGKA captures the basic primitive underlying group messaging schemes. Current solutions including TreeKEM (“Messaging Layer Security” (MLS) IETF RFC 9420) cannot handle concurrent requests while retaining low communication complexity. The exception being CoCoA, which is concurrent while having extremely low communication complexity (in groups of size n and for m concurrent updates the communication per user is log(n), i.e., independent of m). The main downside of CoCoA is that in groups of size n, users might have to do up to log(n) update requests to the server to ensure their (potentially corrupted) key material has been refreshed. In this work we present a “fast healing” concurrent CGKA protocol, named DeCAF, where users will heal after at most log(t) requests, with t being the number of corrupted users. While also suitable for the standard central-server setting, our protocol is particularly interesting for realizing decentralized group messaging, where protocol messages (add, remove, update) are being posted on some append-only data structure rather than sent to a server. In this setting, concurrency is crucial once the rate of requests exceeds, say, the rate at which new blocks are added to a blockchain. In the central-server setting, CoCoA (the only alternative with concurrency, sub-linear communication and basic post-compromise security) enjoys much lower download communication. However, in the decentralized setting – where there is no server which can craft specific messages for different users to reduce their download communication – our protocol significantly outperforms CoCoA. DeCAF heals in fewer epochs (log(t) vs. log(n)) while incurring a similar per epoch per user communication cost.}, author = {Alwen, Joel F and Auerbach, Benedikt and Cueto Noval, Miguel and Klein, Karen and Pascual Perez, Guillermo and Pietrzak, Krzysztof Z}, booktitle = {Security and Cryptography for Networks: 14th International Conference}, editor = {Galdi, Clemente and Phan, Duong Hieu}, isbn = {9783031710728}, issn = {1611-3349}, location = {Amalfi, Italy}, pages = {294–313}, publisher = {Springer Nature}, title = {{DeCAF: Decentralizable CGKA with fast healing}}, doi = {10.1007/978-3-031-71073-5_14}, volume = {14974}, year = {2024}, } @article{18087, abstract = {We present a theory describing the interaction of structured light, such as light carrying orbital angular momentum, with molecules. The light-matter interaction Hamiltonian we derive is expressed through couplings between spherical gradients of the electric field and the (transition) electric multipole moments of a particle of any nontrivial rotation point group. Our model can therefore accommodate an arbitrary complexity of the molecular and electric field structure, and it can be straightforwardly extended to atoms or nanostructures. Applying this framework to rovibrational spectroscopy of molecules, we uncover the general mechanism of angular momentum exchange between the spin and orbital angular momenta of light, molecular rotation, and its center-of-mass motion. We show that the nonzero vorticity of Laguerre-Gaussian beams can strongly enhance certain rovibrational transitions that are considered forbidden in the case of nonhelical light. We discuss the experimental requirements for the observation of these forbidden transitions in state-of-the-art spatially resolved spectroscopy measurements.}, author = {Maslov, Mikhail and Koutentakis, Georgios and Hrast, Mateja and Heckl, Oliver H. and Lemeshko, Mikhail}, issn = {2643-1564}, journal = {Physical Review Research}, number = {3}, publisher = {American Physical Society}, title = {{Theory of angular momentum transfer from light to molecules}}, doi = {10.1103/physrevresearch.6.033277}, volume = {6}, year = {2024}, } @phdthesis{18088, abstract = {Instant messaging applications like Whatsapp, Signal or Telegram have become ubiquitous in today's society. Many of them provide not only end-to-end encryption, but also security guarantees even when the key material gets compromised. These are achieved through frequent key update performed by users. In particular, the compromise of a group key should preserve confidentiality of previously exchanged messages (forward secrecy), and a subsequent key update will ensure security for future ones (post-compromise security). Though great protocols for one-on-one communication have been known for some time, the design of ones that scale efficiently for larger groups while achieving akin security guarantees is a hard problem. A great deal of research has been aimed at this topic, much of it under the umbrella of the Messaging Layer Security (MLS) working group at the IETF. Started in 2018, this joint effort by academics and industry culminated in 2023 with the publication of the first standard for secure group messaging [IETF, RFC9420]. At the core of secure group messaging is a cryptographic primitive termed Continuous Group Key Agreement, or CGKA [Alwen et al. 2021], that essentially allows a changing group of users to agree on a common key with the added functionality security against compromises is achieved by users asynchronously issuing a key update. In this thesis we contribute to the understanding of CGKA across different angles. First, we present a new technique to effect dynamic operations in groups, i.e., add or remove members, that can be more efficient that the one employed by MLS in certain settings. Considering the setting of users belonging to multiple overlapping groups, we then show lowerbounds on the communication cost of constructions that leverage said overlap, at the same time showing protocols that are asymptotically optimal and efficient for practical settings, respectively. Along the way, we show that the communication cost of key updates in MLS is average-cost optimal. An important feature in CGKA protocols, particularly for big groups, is the possibility of executing several group operations concurrently. While later versions of MLS support this, they do at the cost of worsening the communication efficiency of future group operations. In this thesis we introduce two new protocols that permit concurrency without any negative effect on efficiency. Our protocols circumvent previously existing lower bounds by satisfying a new notion of post-compromise security that only asks for security to be re-established after a certain number of key updates have taken place. While this can be slower than MLS in terms of rounds of communication, we show that it leads to more efficient overall communication. Additionally, we introduce a new technique that allows group members to decrease the information they need to store and download, which makes one of our protocols enjoy much lower download cost than any other existing CGKA constructions. }, author = {Pascual Perez, Guillermo}, issn = {2663-337X}, pages = {239}, publisher = {Institute of Science and Technology Austria}, title = {{On the efficiency and security of secure group messaging}}, doi = {10.15479/at:ista:18088}, year = {2024}, } @inproceedings{18097, abstract = {In our companion paper "Tight bounds for the learning of homotopy à la Niyogi, Smale, and Weinberger for subsets of Euclidean spaces and of Riemannian manifolds" we gave optimal bounds (in terms of the two one-sided Hausdorff distances) on a sample P of an input shape 𝒮 (either manifold or general set with positive reach) such that one can infer the homotopy of 𝒮 from the union of balls with some radius centred at P, both in Euclidean space and in a Riemannian manifold of bounded curvature. The construction showing the optimality of the bounds is not straightforward. The purpose of this video is to visualize and thus elucidate said construction in the Euclidean setting.}, author = {Attali, Dominique and Kourimska, Hana and Fillmore, Christopher D and Ghosh, Ishika and Lieutier, Andre and Stephenson, Elizabeth R and Wintraecken, Mathijs}, booktitle = {40th International Symposium on Computational Geometry}, isbn = {9783959773164}, location = {Athens, Greece}, publisher = {Schloss Dagstuhl - Leibniz-Zentrum für Informatik}, title = {{The ultimate frontier: An optimality construction for homotopy inference (media exposition)}}, doi = {10.4230/LIPIcs.SoCG.2024.87}, volume = {293}, year = {2024}, } @article{18107, abstract = {We consider a dilute fully spin-polarized Fermi gas at positive temperature in dimensions d∈{1,2,3} . We show that the pressure of the interacting gas is bounded from below by that of the free gas plus, to leading order, an explicit term of order adρ2+2/d, where a is the p-wave scattering length of the repulsive interaction and ρ is the particle density. The results are valid for a wide range of repulsive interactions, including that of a hard core, and uniform in temperatures at most of the order of the Fermi temperature. A central ingredient in the proof is a rigorous implementation of the fermionic cluster expansion of Gaudin, Gillespie and Ripka (Nucl. Phys. A, 176.2 (1971), pp. 237–260).}, author = {Lauritsen, Asbjørn Bækgaard and Seiringer, Robert}, issn = {2050-5094}, journal = {Forum of Mathematics, Sigma}, publisher = {Cambridge University Press}, title = {{Pressure of a dilute spin-polarized Fermi gas: Lower bound}}, doi = {10.1017/fms.2024.56}, volume = {12}, year = {2024}, } @article{18108, abstract = {Here we announce the construction and properties of a big commutative subalgebra of the Kirillov algebra attached to a finite dimensional irreducible representation of a complex semisimple Lie group. They are commutative finite flat algebras over the cohomology of the classifying space of the group. They are isomorphic with the equivariant intersection cohomology of affine Schubert varieties, endowing the latter with a new ring structure. Study of the finer aspects of the structure of the big algebras will also furnish the stalks of the intersection cohomology with ring structure, thus ringifying Lusztig’s q-weight multiplicity polynomials i.e., certain affine Kazhdan–Lusztig polynomials.}, author = {Hausel, Tamás}, issn = {1091-6490}, journal = {Proceedings of the National Academy of Sciences of the United States of America}, number = {38}, publisher = {National Academy of Sciences}, title = {{Commutative avatars of representations of semisimple Lie groups}}, doi = {10.1073/pnas.2319341121}, volume = {121}, year = {2024}, } @article{18109, abstract = {Venous thromboembolism (VTE) is a common, deadly disease with an increasing incidence despite preventive efforts. Clinical observations have associated elevated antibody concentrations or antibody-based therapies with thrombotic events. However, how antibodies contribute to thrombosis is unknown. Here, we show that reduced blood flow enabled immunoglobulin M (IgM) to bind to FcμR and the polymeric immunoglobulin receptor (pIgR), initiating endothelial activation and platelet recruitment. Subsequently, the procoagulant surface of activated platelets accommodated antigen- and FcγR-independent IgG deposition. This leads to classical complement activation, setting in motion a prothrombotic vicious circle. Key elements of this mechanism were present in humans in the setting of venous stasis as well as in the dysregulated immunothrombosis of COVID-19. This antibody-driven thrombosis can be prevented by pharmacologically targeting complement. Hence, our results uncover antibodies as previously unrecognized central regulators of thrombosis. These findings carry relevance for therapeutic application of antibodies and open innovative avenues to target thrombosis without compromising hemostasis.}, author = {Stark, Konstantin and Kilani, Badr and Stockhausen, Sven and Busse, Johanna and Schubert, Irene and Tran, Thuy Duong and Gärtner, Florian R and Leunig, Alexander and Pekayvaz, Kami and Nicolai, Leo and Fumagalli, Valeria and Stermann, Julia and Stephan, Felix and David, Christian and Müller, Martin B. and Heyman, Birgitta and Lux, Anja and Da Palma Guerreiro, Alexandra and Frenzel, Lukas P. and Schmidt, Christoph Q. and Dopler, Arthur and Moser, Markus and Chandraratne, Sue and Von Brühl, Marie Luise and Lorenz, Michael and Korff, Thomas and Rudelius, Martina and Popp, Oliver and Kirchner, Marieluise and Mertins, Philipp and Nimmerjahn, Falk and Iannacone, Matteo and Sperandio, Markus and Engelmann, Bernd and Verschoor, Admar and Massberg, Steffen}, issn = {1097-4180}, journal = {Immunity}, number = {9}, pages = {2140--2156}, publisher = {Elsevier}, title = {{Antibodies and complement are key drivers of thrombosis}}, doi = {10.1016/j.immuni.2024.08.007}, volume = {57}, year = {2024}, } @article{18110, abstract = {We study a chaotic particle-conserving kinetically constrained model, with a single parameter which allows us to break reflection symmetry. Through extensive numerical simulations we find that the domain wall state shows a variety of dynamical behaviors from localization all the way to ballistic transport, depending on the value of the reflection breaking parameter. Surprisingly, such anomalous behavior is not mirrored in infinite-temperature dynamics, which appear to scale diffusively, in line with expectations for generic interacting models. However, studying the particle density gradient, we show that the lack of reflection symmetry affects infinite-temperature dynamics, resulting in an asymmetric dynamical structure factor. This is in disagreement with normal diffusion and suggests that the model may also exhibit anomalous dynamics at infinite temperature in the thermodynamic limit. Finally, we observe low-entangled eigenstates in the spectrum of the model, a telltale sign of quantum many-body scars.}, author = {Brighi, Pietro and Ljubotina, Marko}, issn = {2469-9969}, journal = {Physical Review B}, number = {10}, publisher = {American Physical Society}, title = {{Anomalous transport in the kinetically constrained quantum East-West model}}, doi = {10.1103/PhysRevB.110.L100304}, volume = {110}, year = {2024}, } @article{18111, abstract = {Observations of tidal disruption events (TDEs) show signs of nitrogen enrichment reminiscent of other astrophysical sources such as active galactic nuclei and star-forming galaxies. Given that TDEs probe the gas from a single star, it is possible to test whether the observed enrichment is consistent with expectations from the CNO cycle by looking at the observed nitrogen/carbon (N/C) abundance ratios. Given that ≈20% of solar-mass stars (and an even larger fraction of more massive stars) live in close binaries, it is worthwhile to also consider what TDEs from stars influenced by binary evolution would look like. We show here that TDEs from stars stripped of their hydrogen-rich (and nitrogen-poor) envelopes through previous binary-induced mass loss can produce much higher observable N/C enhancements than even TDEs from massive stars. Additionally, we predict that the time dependence of the N/C abundance ratio in the mass fallback rate of stripped stars will follow the inverse behavior of main-sequence stars, enabling a more accurate characterization of the disrupted star.}, author = {Mockler, Brenna and Gallegos-Garcia, Monica and Götberg, Ylva Louise Linsdotter and Miller, Jon M. and Ramirez-Ruiz, Enrico}, issn = {2041-8213}, journal = {Astrophysical Journal Letters}, number = {1}, publisher = {IOP Publishing}, title = {{Tidal disruption events from stripped stars}}, doi = {10.3847/2041-8213/ad6c34}, volume = {973}, year = {2024}, } @inproceedings{18113, abstract = {The emergence of accurate open large language models (LLMs) has led to a race towards performant quantization techniques which can enable their execution on end-user devices. In this paper, we revisit the problem of “extreme” LLM compression—defined as targeting extremely low bit counts, such as 2 to 3 bits per parameter—from the point of view of classic methods in Multi-Codebook Quantization (MCQ). Our algorithm, called AQLM, generalizes the classic Additive Quantization (AQ) approach for information retrieval to advance the state-of-the-art in LLM compression, via two innovations: 1) learned additive quantization of weight matrices in input-adaptive fashion, and 2) joint optimization of codebook parameters across each transformer blocks. Broadly, AQLM is the first scheme that is Pareto optimal in terms of accuracy-vs-model-size when compressing to less than 3 bits per parameter, and significantly improves upon all known schemes in the extreme compression (2bit) regime. In addition, AQLM is practical: we provide fast GPU and CPU implementations of AQLM for token generation, which enable us to match or outperform optimized FP16 implementations for speed, while executing in a much smaller memory footprint.}, author = {Egiazarian, Vage and Panferov, Andrei and Kuznedelev, Denis and Frantar, Elias and Babenko, Artem and Alistarh, Dan-Adrian}, booktitle = {Proceedings of the 41st International Conference on Machine Learning}, issn = {2640-3498}, location = {Vienna, Austria}, pages = {12284--12303}, publisher = {ML Research Press}, title = {{Extreme compression of large language models via additive quantization}}, volume = {235}, year = {2024}, } @inproceedings{18114, abstract = {This paper presents Mechanistic Neural Networks, a neural network design for machine learning applications in the sciences. It incorporates a new Mechanistic Block in standard architectures to explicitly learn governing differential equations as representations, revealing the underlying dynamics of data and enhancing interpretability and efficiency in data modeling. Central to our approach is a novel Relaxed Linear Programming Solver (NeuRLP) inspired by a technique that reduces solving linear ODEs to solving linear programs. This integrates well with neural networks and surpasses the limitations of traditional ODE solvers enabling scalable GPU parallel processing. Overall, Mechanistic Neural Networks demonstrate their versatility for scientific machine learning applications, adeptly managing tasks from equation discovery to dynamic systems modeling. We prove their comprehensive capabilities in analyzing and interpreting complex scientific data across various applications, showing significant performance against specialized state-of-the-art methods. Source code is available at https://github.com/alpz/mech-nn.}, author = {Pervez, Adeel A and Locatello, Francesco and Gavves, Efstratios}, booktitle = {Proceedings of the 41st International Conference on Machine Learning}, issn = {2640-3498}, location = {Vienna, Austria}, pages = {40484--40501}, publisher = {ML Research Press}, title = {{Mechanistic neural networks for scientific machine learning}}, volume = {235}, year = {2024}, } @inproceedings{18115, abstract = {We study the data selection problem, whose aim is to select a small representative subset of data that can be used to efficiently train a machine learning model. We present a new data selection approach based on k-means clustering and sensitivity sampling. Assuming access to an embedding representation of the data with respect to which the model loss is Holder continuous, our approach provably allows selecting a set of “typical” k+1/ε2 elements whose average loss corresponds to the average loss of the whole dataset, up to a multiplicative (1±ε) factor and an additive ελΦk, where Φk represents the k-means cost for the input embeddings and λ is the Holder constant. We furthermore demonstrate the performance and scalability of our approach on fine-tuning foundation models and show that it outperforms state-of-the-art methods. We also show how it can be applied on linear regression, leading to a new sampling strategy that surprisingly matches the performance of leverage score sampling, while being conceptually simpler and more scalable.}, author = {Axiotis, Kyriakos and Cohen-Addad, Vincent and Henzinger, Monika H and Jerome, Sammy and Mirrokni, Vahab and Saulpic, David and Woodruff, David P. and Wunder, Michael}, booktitle = {Proceedings of the 41st International Conference on Machine Learning}, issn = {2640-3498}, location = {Vienna, Austria}, pages = {2086--2107}, publisher = {ML Research Press}, title = {{Data-efficient learning via clustering-based sensitivity sampling: Foundation models and beyond}}, volume = {235}, year = {2024}, } @inproceedings{18116, abstract = {As a staple of data analysis and unsupervised learning, the problem of private clustering has been widely studied, under various privacy models. Centralized differential privacy is the first of them, and the problem has also been studied for the local and the shuffle variation. In each case, the goal is to design an algorithm that computes privately a clustering, with the smallest possible error. The study of each variation gave rise to new algorithm: the landscape of private clustering algorithm is therefore quite intricate. In this paper, we show that a 20 year-old algorithm can be slightly modified to work for any of those models. This provides a unified picture: while matching almost all previously known results, it allows us to improve some of them, and extend to a new privacy model, the continual observation setting, where the input is changing over time and the algorithm must output a new solution at each time step.}, author = {La Tour, Max Dupré and Henzinger, Monika H and Saulpic, David}, booktitle = {Proceedings of the 41st International Conference on Machine Learning}, issn = {2640-3498}, location = {Vienna, Austria}, pages = {12046--12086}, publisher = {ML Research Press}, title = {{Making old things new: A unified algorithm for differentially private clustering}}, volume = {235}, year = {2024}, } @inproceedings{18117, abstract = {We investigate parameter-efficient fine-tuning (PEFT) methods that can provide good accuracy under limited computational and memory budgets in the context of large language models (LLMs). We present a new PEFT method called Robust Adaptation (RoSA) inspired by robust principal component analysis that jointly trains low-rank and highly-sparse components on top of a set of fixed pretrained weights to efficiently approximate the performance of a full-fine-tuning (FFT) solution. Across a series of challenging generative tasks such as grade-school math and SQL query generation, which require fine-tuning for good performance, we show that RoSA outperforms LoRA, pure sparse fine-tuning, and alternative hybrid methods at the same parameter budget, and can even recover the performance of FFT on some tasks. We provide system support for RoSA to complement the training algorithm, specifically in the form of sparse GPU kernels which enable memory- and computationally-efficient training, and show that it is also compatible with low-precision base weights, resulting in the first joint representation combining quantization, low-rank and sparse approximations. Our code is available at https://github.com/IST-DASLab/RoSA.}, author = {Nikdan, Mahdi and Tabesh, Soroush and Crncevic, Elvir and Alistarh, Dan-Adrian}, booktitle = {Proceedings of the 41st International Conference on Machine Learning}, issn = {2640-3498}, location = {Vienna, Austria}, pages = {38187--38206}, publisher = {ML Research Press}, title = {{RoSA: Accurate parameter-efficient fine-tuning via robust adaptation}}, volume = {235}, year = {2024}, } @inproceedings{18118, abstract = {We introduce a new framework for studying meta-learning methods using PAC-Bayesian theory. Its main advantage over previous work is that it allows for more flexibility in how the transfer of knowledge between tasks is realized. For previous approaches, this could only happen indirectly, by means of learning prior distributions over models. In contrast, the new generalization bounds that we prove express the process of meta-learning much more directly as learning the learning algorithm that should be used for future tasks. The flexibility of our framework makes it suitable to analyze a wide range of meta-learning mechanisms and even design new mechanisms. Other than our theoretical contributions we also show empirically that our framework improves the prediction quality in practical meta-learning mechanisms.}, author = {Zakerinia, Hossein and Behjati, Amin and Lampert, Christoph}, booktitle = {Proceedings of the 41st International Conference on Machine Learning}, issn = {2640-3498}, location = {Vienna, Austria}, pages = {58122--58139}, publisher = {ML Research Press}, title = {{More flexible PAC-Bayesian meta-learning by learning learning algorithms}}, volume = {235}, year = {2024}, } @inproceedings{18120, abstract = {In practice, training using federated learning can be orders of magnitude slower than standard centralized training. This severely limits the amount of experimentation and tuning that can be done, making it challenging to obtain good performance on a given task. Server-side proxy data can be used to run training simulations, for instance for hyperparameter tuning. This can greatly speed up the training pipeline by reducing the number of tuning runs to be performed overall on the true clients. However, it is challenging to ensure that these simulations accurately reflect the dynamics of the real federated training. In particular, the proxy data used for simulations often comes as a single centralized dataset without a partition into distinct clients, and partitioning this data in a naive way can lead to simulations that poorly reflect real federated training. In this paper we address the challenge of how to partition centralized data in a way that reflects the statistical heterogeneity of the true federated clients. We propose a fully federated, theoretically justified, algorithm that efficiently learns the distribution of the true clients and observe improved server-side simulations when using the inferred distribution to create simulated clients from the centralized data.}, author = {Scott, Jonathan A and Cahill, Áine}, booktitle = {Proceedings of the 41st International Conference on Machine Learning}, issn = {2640-3498}, location = {Vienna, Austria}, pages = {44012--44037}, publisher = {ML Research Press}, title = {{Improved modelling of federated datasets using mixtures-of-Dirichlet-multinomials}}, volume = {235}, year = {2024}, } @phdthesis{18135, abstract = {This thesis consists of two separate parts. In the first part we consider a dilute Fermi gas interacting through a repulsive interaction in dimensions $d=1,2,3$. Our focus is mostly on the physically most relevant dimension $d=3$ and the setting of a spin-polarized (equivalently spinless) gas, where the Pauli exclusion principle plays a key role. We show that, at zero temperature, the ground state energy density of the interacting spin-polarized gas differs (to leading order) from that of the free (i.e. non-interacting) gas by a term of order $a_p^d\rho^{2+2/d}$ with $a_p$ the $p$-wave scattering length of the repulsive interaction and $\rho$ the density. Further, we extend this to positive temperature and show that the pressure of an interacting spin-polarized gas differs from that of the free gas by a now temperature dependent term, again of order $a_p^d\rho^{2+2/d}$. Lastly, we consider the setting of a spin-$\frac{1}{2}$ Fermi gas in $d=3$ dimensions and show that here, as an upper bound, the ground state energy density differs from that of the free system by a term of order $a_s \rho^2$ with an error smaller than $a_s \rho^2 (a_s\rho^{1/3})^{1-\eps}$ for any $\eps > 0$, where $a_s$ is the $s$-wave scattering length of the repulsive interaction. These asymptotic formulas complement the similar formulas in the literature for the dilute Bose and spin-$\frac{1}{2}$ Fermi gas, where the ground state energies or pressures differ from that of the corresponding free systems by a term of order $a_s \rho^2$ in dimension $d=3$. In the spin-polarized setting, the corrections, of order $a_p^3\rho^{8/3}$ in dimension $d=3$, are thus much smaller and requires a more delicate analysis. In the second part of the thesis we consider the Bardeen--Cooper--Schrieffer (BCS) theory of superconductivity and in particular its associated critical temperature and energy gap. We prove that the ratio of the zero-temperature energy gap and critical temperature $\Xi(T=0)/T_c$ approaches a universal constant $\pi e^{-\gamma}\approx 1.76$ in both the limit of high density in dimension $d=3$ and in the limit of weak coupling in dimensions $d=1,2$. This complements the proofs in the literature of this universal behaviour in the limit of weak coupling or low density in dimension $d=3$. Secondly, we prove that the ratio of the energy gap at positive temperature and critical temperature $\Xi(T)/T_c$ approaches a universal function of the relative temperature $T/T_c$ in the limit of weak coupling in dimensions $d=1,2,3$.}, author = {Lauritsen, Asbjørn Bækgaard}, isbn = {978-3-99078-042-8}, issn = {2663-337X}, pages = {353}, publisher = {Institute of Science and Technology Austria}, title = {{Energies of dilute Fermi gases and universalities in BCS theory}}, doi = {10.15479/at:ista:18135}, year = {2024}, } @article{18153, abstract = {Attention supports decision making by selecting the features that are relevant for decisions. Selective enhancement of the relevant features and inhibition of distractors has been proposed as potential neural mechanisms driving this selection process. Yet, how attention operates when relevance cannot be directly determined, and the attention signal needs to be internally constructed is less understood. Here we recorded from populations of neurons in the anterior cingulate cortex (ACC) of mice in an attention-shifting task where relevance of stimulus modalities changed across blocks of trials. In contrast with V1 recordings, decoding of the irrelevant modality gradually declined in ACC after an initial transient. Our analytical proof and a recurrent neural network model of the task revealed mutually inhibiting connections that produced context-gated suppression as observed in mice. Using this RNN model we predicted a correlation between contextual modulation of individual neurons and their stimulus drive, which we confirmed in ACC but not in V1.}, author = {Hajnal, Márton Albert and Tran, Duy and Szabó, Zsombor and Albert, Andrea and Safaryan, Karen and Einstein, Michael and Vallejo Martelo, Mauricio and Polack, Pierre-Olivier and Golshani, Peyman and Orbán, Gergő}, issn = {2041-1723}, journal = {Nature Communications}, publisher = {Springer Nature}, title = {{Shifts in attention drive context-dependent subspace encoding in anterior cingulate cortex in mice during decision making}}, doi = {10.1038/s41467-024-49845-2}, volume = {15}, year = {2024}, } @inproceedings{18155, abstract = {We study the classical problem of verifying programs with respect to formal specifications given in the linear temporal logic (LTL). We first present novel sound and complete witnesses for LTL verification over imperative programs. Our witnesses are applicable to both verification (proving) and refutation (finding bugs) settings. We then consider LTL formulas in which atomic propositions can be polynomial constraints and turn our focus to polynomial arithmetic programs, i.e. programs in which every assignment and guard consists only of polynomial expressions. For this setting, we provide an efficient algorithm to automatically synthesize such LTL witnesses. Our synthesis procedure is both sound and semi-complete. Finally, we present experimental results demonstrating the effectiveness of our approach and that it can handle programs which were beyond the reach of previous state-of-the-art tools.}, author = {Chatterjee, Krishnendu and Goharshady, Amir Kafshdar and Goharshady, Ehsan and Karrabi, Mehrdad and Zikelic, Dorde}, booktitle = {Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)}, isbn = {9783031711619}, issn = {1611-3349}, location = {Milan, Italy}, pages = {600--619}, publisher = {Springer Nature}, title = {{Sound and complete witnesses for template-based verification of LTL properties on polynomial programs}}, doi = {10.1007/978-3-031-71162-6_31}, volume = {14933}, year = {2024}, } @inproceedings{18156, abstract = {Privately counting distinct elements in a stream is a fundamental data analysis problem with many applications in machine learning. In the turnstile model, Jain et al. [NeurIPS2023] initiated the study of this problem parameterized by the maximum flippancy of any element, i.e., the number of times that the count of an element changes from 0 to above 0 or vice versa. They give an item-level (ε,δ)-differentially private algorithm whose additive error is tight with respect to that parameterization. In this work, we show that a very simple algorithm based on the sparse vector technique achieves a tight additive error for item-level (ε,δ)-differential privacy and item-level ε-differential privacy with regards to a different parameterization, namely the sum of all flippancies. Our second result is a bound which shows that for a large class of algorithms, including all existing differentially private algorithms for this problem, the lower bound from item-level differential privacy extends to event-level differential privacy. This partially answers an open question by Jain et al. [NeurIPS2023].}, author = {Henzinger, Monika H and Sricharan, A. R. and Steiner, Teresa Anna}, booktitle = {International Conference on Approximation Algorithms for Combinatorial Optimization Problems }, isbn = {9783959773485}, issn = {1868-8969}, location = {London, United Kingdom}, publisher = {Schloss Dagstuhl - Leibniz-Zentrum für Informatik}, title = {{Private counting of distinct elements in the turnstile model and extensions}}, doi = {10.4230/LIPIcs.APPROX/RANDOM.2024.40}, volume = {317}, year = {2024}, }