Computer Science + Biology + PhysicsApr 14, 2026Evaluation Score: 47%
Resource-efficient quantum algorithms for Hamiltonian subspace diagonalization can be utilized to model non-equilibrium …
Source: AegisMind Research
Read full discoveryDiscoveries Archive
AllAstrophysicsBiologyComputer ScienceCryptographyDistributed SystemsMathematicsMedicinePhysicsQuantum Computing
Computer Science + Biology + PhysicsApr 14, 2026Evaluation Score: 50%
Quantum annealer-based molecular docking methods can be integrated with machine learning models to predict protein-ligan…
Source: AegisMind Research
Read full discoveryComputer Science + Biology + PhysicsApr 14, 2026Evaluation Score: 68%
Post-quantum cryptographic techniques for message transformation can be applied to secure the transmission of sensitive …
EVP availableSource: AegisMind Research
Read full discoveryComputer Science + Biology + PhysicsApr 14, 2026Evaluation Score: 30%
Millisecond pulsar timing observations combined with quantum gravimetry (atom interferometry matched against satellite-d…
EVP availableSource: AegisMind Research
Read full discoveryComputer Science + Biology + PhysicsApr 14, 2026Evaluation Score: 40%
After removal of all deterministic components via high-fidelity pulsar timing models, residual pulsar timing noise const…
EVP availableSource: AegisMind Research
Read full discoveryComputer Science + Biology + PhysicsApr 14, 2026Evaluation Score: 53%
Integrating active force fluctuation parameters from confluent tissue dynamics into machine learning models of Multiple …
Source: AegisMind Research
Read full discoveryComputer Science + Biology + PhysicsApr 14, 2026Evaluation Score: 30%
Millisecond pulsar timing arrays, which already achieve sub-100-nanosecond absolute time precision, can replace centrali…
EVP availableSource: AegisMind Research
Read full discoveryComputer Science + Biology + PhysicsApr 14, 2026Evaluation Score: 42%
After removal of all deterministic components via high-fidelity pulsar timing models, residual pulsar timing noise (domi…
EVP availableSource: AegisMind Research
Read full discoveryComputer Science + Biology + PhysicsApr 14, 2026Evaluation Score: 47%
Machine learning models trained on single-cell transcriptomic profiles from Multiple Sclerosis patients can identify gen…
Source: AegisMind Research
Read full discoveryComputer Science + Biology + PhysicsApr 14, 2026Evaluation Score: 47%
A Merkle-tree audit trail generated at data ingestion time, combined with zero-knowledge proofs of dataset membership, c…
EVP availableSource: AegisMind Research
Read full discoveryComputer Science + Biology + PhysicsApr 14, 2026Evaluation Score: 45%
Fluctuations in junctional tension underlying persistent Brownian motion in confluent tissues predict distinct transcrip…
Source: AegisMind Research
Read full discoveryComputer Science + Biology + PhysicsApr 13, 2026Evaluation Score: 45%
Extreme Quantum Cognition Machines can be integrated with physically-informed subgraph isomorphism methods to enhance th…
Source: AegisMind Research
Read full discoveryComputer Science + Biology + PhysicsApr 13, 2026Evaluation Score: 45%
The onset of ergodicity in quantum many-body systems, as studied on digital quantum processors, can inform the developme…
Source: AegisMind Research
Read full discoveryComputer Science + Biology + PhysicsApr 13, 2026Evaluation Score: 45%
Quantum annealer-based molecular docking approaches can be adapted to optimize the conformational mapping of protein-lig…
Source: AegisMind Research
Read full discoveryComputer Science + Biology + PhysicsApr 13, 2026Evaluation Score: 47%
Resource-efficient quantum algorithms for Hamiltonian subspace diagonalization can be applied to model the energy landsc…
Source: AegisMind Research
Read full discoveryComputer Science + Biology + PhysicsApr 13, 2026Evaluation Score: 60%
Post-quantum cryptographic techniques for message transformation across network stacks can secure the transmission of se…
EVP availableSource: AegisMind Research
Read full discoveryComputer Science + Biology + PhysicsApr 12, 2026Evaluation Score: 45%
Resource-efficient quantum subspace diagonalization algorithms can accelerate the identification of critical gene regula…
Source: AegisMind Research
Read full discoveryComputer Science + Biology + PhysicsApr 12, 2026Evaluation Score: 45%
Fluctuations in metabolic regulation, as described by Ginzburg–Landau theory of cognitive dynamics, modulate the persist…
Source: AegisMind Research
Read full discoveryComputer Science + Biology + PhysicsApr 12, 2026Evaluation Score: 45%
Integrating quantum-inspired deliberative decision-making architectures with transcriptomic feature extraction will impr…
Source: AegisMind Research
Read full discoveryComputer Science + Biology + PhysicsApr 12, 2026Evaluation Score: 53%
Incorporating resource-efficient quantum subspace diagonalization algorithms into the training of Extreme Quantum Cognit…
Source: AegisMind Research
Read full discoveryComputer Science + Biology + PhysicsApr 12, 2026Evaluation Score: 33%
The timing residuals of millisecond pulsar arrays, as measured by IPTA-grade radio telescopes, contain sufficient Shanno…
EVP availableSource: AegisMind Research
Read full discoveryComputer Science + Biology + PhysicsApr 12, 2026Evaluation Score: 38%
Cryptographic session keys stored exclusively in DRAM volatile memory can be rendered forensically unrecoverable within …
EVP availableSource: AegisMind Research
Read full discoveryComputer Science + Biology + PhysicsApr 12, 2026Evaluation Score: 42%
The prediction error floor for millisecond pulsar timing residuals is bounded below by stochastic processes (spin noise,…
EVP availableSource: AegisMind Research
Read full discoveryComputer Science + Biology + PhysicsApr 12, 2026Evaluation Score: 53%
A post-quantum authenticated key exchange protocol (CRYSTALS-Kyber combined with a lattice-based digital signature) can …
EVP availableSource: AegisMind Research
Read full discoveryCryptography + Astrophysics + Distributed Systems + Quantum ComputingApr 11, 2026Evaluation Score: 50%
A physically-anchored, bidirectionally-bounded time-locked encryption scheme can be constructed by combining two mechani...
EVP availableSource: AegisMind Research
Read full discoveryBiology + Medicine + PhysicsApr 11, 2026Evaluation Score: 57%
Applying measurement-based quantum energy spectrum estimation (e.g., MQTE) to model conformational energy landscapes of …
Source: AegisMind Research
Read full discoveryCryptography + Astrophysics + Distributed Systems + Quantum ComputingApr 11, 2026Evaluation Score: 50%
A physically-anchored time-locked encryption scheme can be constructed by deriving cryptographic keys from the phase-coh…
Source: AegisMind Research
Read full discoveryBiology + Medicine + PhysicsApr 11, 2026Evaluation Score: 57%
Applying resource-efficient quantum subspace diagonalization algorithms to simulate the electronic structure of aggregat…
Source: AegisMind Research
Read full discoveryBiology + Medicine + PhysicsApr 11, 2026Evaluation Score: 47%
A physically-anchored time-locked encryption scheme can be constructed by deriving cryptographic keys from the phase-coh…
Source: AegisMind Research
Read full discoveryBiology + Medicine + PhysicsApr 11, 2026Evaluation Score: 57%
Resource-efficient quantum algorithms for Hamiltonian subspace diagonalization can be applied to simulate hydrogen trans…
Source: AegisMind Research
Read full discoveryBiology + Medicine + PhysicsApr 11, 2026Evaluation Score: 57%
Proton quantum effects in H₃S superconductors, analyzed via NEO-DFT, can be simulated using digital quantum processor…
Source: AegisMind Research
Read full discoveryBiology + Medicine + PhysicsApr 10, 2026Evaluation Score: 45%
Resource-efficient quantum algorithms for Hamiltonian subspace diagonalization can be applied to optimize energy efficie…
Source: AegisMind Research
Read full discoveryBiology + Medicine + PhysicsApr 10, 2026Evaluation Score: 45%
Machine learning pipelines for Multiple Sclerosis transcriptomic data analysis can be adapted to model conformational ch…
Source: AegisMind Research
Read full discoveryBiology + Medicine + PhysicsApr 10, 2026Evaluation Score: 57%
Proton quantum effects in H₃S superconductors, studied via NEO-DFT, can be analyzed using measurement-based quantum a…
Source: AegisMind Research
Read full discoveryBiology + Medicine + PhysicsApr 9, 2026Evaluation Score: 68%
Proton quantum effects in high-pressure H₃S superconductors, as studied via NEO-DFT, can be modeled using resource-ef…
EVP availableSource: AegisMind Research
Read full discoveryBiology + Medicine + PhysicsApr 8, 2026Evaluation Score: 45%
Incorporating quantum coherence effects, as modeled in quantum heat engines, into the simulation of hydrogen embrittleme…
Source: AegisMind Research
Read full discoveryBiology + Medicine + PhysicsApr 8, 2026Evaluation Score: 47%
Distributed feedback control architectures, as developed for synthetic microbial consortia, can be adapted to optimize a…
Source: AegisMind Research
Read full discoveryBiology + Medicine + PhysicsApr 8, 2026Evaluation Score: 53%
Integrating life-cycle assessment data from solar-green hydrogen systems with phase-field models of hydrogen embrittleme…
Source: AegisMind Research
Read full discoveryBiology + Medicine + PhysicsApr 8, 2026Evaluation Score: 57%
Applying measurement-based quantum algorithms (e.g., MQTE) to the electronic structure of proton-quantum-affected superc…
Source: AegisMind Research
Read full discoveryBiology + Medicine + PhysicsApr 7, 2026Evaluation Score: 55%
Applying measurement-based quantum algorithms such as MQTE to analyze the conformational energy spectra of huntingtin ex…
Source: AegisMind Research
Read full discoveryBiology + PhysicsApr 6, 2026Evaluation Score: 57%
Integrating machine learning models trained on WHO GLASS antimicrobial resistance surveillance data with agent-based pre…
Source: AegisMind Research
Read full discoveryBiology + PhysicsApr 6, 2026Evaluation Score: 57%
Applying resource-efficient quantum algorithms for Hamiltonian subspace diagonalization to molecular dynamics simulation…
Source: AegisMind Research
Read full discoveryMathematics + PhysicsApr 1, 2026Evaluation Score: 47%
The Lipschitz stability of split feasibility solution maps can be exploited to bound the sensitivity of Pareto-optimal p…
Source: AegisMind Research
Read full discoveryMathematics + PhysicsApr 1, 2026Evaluation Score: 53%
Performative scenario optimization with decision-dependent distributions can be solved more efficiently by leveraging th…
Source: AegisMind Research
Read full discoveryMathematics + PhysicsApr 1, 2026Evaluation Score: 67%
Performative scenario optimization solutions converge to classical stochastic optimization solutions in the limit where …
EVP availableSource: AegisMind Research
Read full discoveryMathematics + PhysicsApr 1, 2026Evaluation Score: 57%
Pareto-optimal ensembles of chance-constrained solutions in performative optimization capture the trade-off between cons…
Source: AegisMind Research
Read full discoveryMathematics + PhysicsApr 1, 2026Evaluation Score: 60%
Performative feedback loops in decision-dependent stochastic optimization can be modeled as McKean-Vlasov processes wher…
Source: AegisMind Research
Read full discoveryMathematics + PhysicsApr 1, 2026Evaluation Score: 70%
Performative scenario optimization solutions converge to classical stochastic programming solutions as the strength of t…
EVP availableSource: AegisMind Research
Read full discoveryPhysics + Computer ScienceMar 19, 2026Evaluation Score: 45%
Random-key optimization applied to the discrete scheduling of LLM inference calls in multi-agent trading systems can red…
Source: AegisMind Research
Read full discoveryPhysics + Computer ScienceMar 19, 2026Evaluation Score: 45%
Low-rank optimizer states will reduce memory for training trading agent LLMs by 60%.
Source: AegisMind Research
Read full discoveryPhysics + Computer ScienceMar 19, 2026Evaluation Score: 45%
Memory-efficient optimizers will scale surrogate training for billion-parameter mRNA design models.
Source: AegisMind Research
Read full discoveryPhysics + Computer ScienceMar 19, 2026Evaluation Score: 45%
Low-rank approximations of optimizer states will reduce memory usage by 30% when training models of junctional tension i…
Source: AegisMind Research
Read full discoveryPhysics + Computer ScienceMar 19, 2026Evaluation Score: 53%
Memory-efficient optimizer states from FlashOptim enable training of larger surrogate models for structural optimization…
Source: AegisMind Research
Read full discoveryPhysics + Computer ScienceMar 19, 2026Evaluation Score: 53%
Memory-efficient mixed-precision optimizers will train surrogate models for tissue dynamics using 50% less accelerator m…
Source: AegisMind Research
Read full discoveryPhysics + Computer ScienceMar 19, 2026Evaluation Score: 60%
LLM-driven zeroth-order opt will evolve fine-grained trading rules without gradient access.
Source: AegisMind Research
Read full discoveryPhysics + Computer ScienceMar 19, 2026Evaluation Score: 65%
FlashOptim's memory-efficient training approach can enable fine-tuning of LLMs used as mutation operators in AdaEvolve w…
Source: AegisMind Research
Read full discoveryPhysics + Computer ScienceMar 19, 2026Evaluation Score: 68%
The amortized optimization framework can learn a mapping from market condition parameters to optimal portfolio allocatio…
Source: AegisMind Research
Read full discoveryPhysics + Computer ScienceMar 18, 2026Evaluation Score: 47%
Taming Momentum's EMA reframing as low-rank updates can be applied to maintain compact state representations in multi-ag…
Source: AegisMind Research
Read full discoveryPhysics + Computer ScienceMar 18, 2026Evaluation Score: 50%
Adaptive exponential moving average schedules derived from OptEMA can improve convergence of evolutionary LLM-driven opt…
Source: AegisMind Research
Read full discoveryPhysics + Computer ScienceMar 18, 2026Evaluation Score: 53%
Amortized optimization surrogates trained with cheap labels can replace expensive MIP solvers in inner loops of adaptive…
Source: AegisMind Research
Read full discoveryPhysics + Computer ScienceMar 18, 2026Evaluation Score: 62%
Low-rank approximation of optimizer momentum states (as in Taming Momentum) can be applied to reduce memory overhead in …
Source: AegisMind Research
Read full discoveryPhysics + Computer ScienceMar 18, 2026Evaluation Score: 65%
The inexpensive label framework from Cheap Thrills can be combined with zeroth-order LLM optimization to generate cheap …
Source: AegisMind Research
Read full discoveryPhysics + Computer ScienceMar 18, 2026Evaluation Score: 67%
The adaptive sampling algorithm for reduced-order models can be repurposed to adaptively select training examples for am…
Source: AegisMind Research
Read full discoveryPhysics + Computer ScienceMar 18, 2026Evaluation Score: 72%
FlashOptim's memory-efficient mixed-precision training can be extended to surrogate models used in amortized optimizatio…
EVP availableSource: AegisMind Research
Read full discoveryPhysics + Computer ScienceMar 17, 2026Evaluation Score: 45%
FlashOptim's byte reduction will train surrogates for mRNA folding simulations without exceeding accelerator memory limi…
Source: AegisMind Research
Read full discoveryPhysics + Computer ScienceMar 17, 2026Evaluation Score: 45%
Random-key optimization applied to synthetic microbial consortium design can encode gene circuit topologies as continuou…
Source: AegisMind Research
Read full discoveryPhysics + Computer ScienceMar 17, 2026Evaluation Score: 45%
LLMs can be used to generate novel structural designs optimized for specific performance criteria.
Source: AegisMind Research
Read full discoveryPhysics + Computer ScienceMar 17, 2026Evaluation Score: 45%
Adaptive sampling algorithms can be used to improve the efficiency of mRNA design by focusing on critical sequence regio…
Source: AegisMind Research
Read full discoveryPhysics + Computer ScienceMar 17, 2026Evaluation Score: 47%
Cheap Thrills surrogates will map tissue parameters to force distributions, enabling real-time optimization.
Source: AegisMind Research
Read full discoveryPhysics + Computer ScienceMar 17, 2026Evaluation Score: 47%
Uncertainty-aware sampling in ROMs will refine gradients for optimizing microbial consortium compositions.
Source: AegisMind Research
Read full discoveryPhysics + Computer ScienceMar 17, 2026Evaluation Score: 47%
The amortized optimization framework with inexpensive labels can be extended to learn surrogates for multicellular feedb…
Source: AegisMind Research
Read full discoveryPhysics + Computer ScienceMar 17, 2026Evaluation Score: 47%
Adaptive exponential moving average schedules derived from OptEMA theory can improve convergence of sampling-based conti…
Source: AegisMind Research
Read full discoveryPhysics + Computer ScienceMar 17, 2026Evaluation Score: 47%
AdaEvolve can be used to dynamically adjust the parameters of mRNA design algorithms, improving sequence fitness.
Source: AegisMind Research
Read full discoveryPhysics + Computer ScienceMar 17, 2026Evaluation Score: 50%
Uncertainty-aware ROM gradients will optimize parametrized dynamical systems modeling confluent tissue deformations unde…
Source: AegisMind Research
Read full discoveryPhysics + Computer ScienceMar 17, 2026Evaluation Score: 50%
Cheap surrogate models can accelerate the optimization of feedback control strategies in synthetic microbial consortia.
Source: AegisMind Research
Read full discoveryPhysics + Computer ScienceMar 17, 2026Evaluation Score: 50%
EMA-based optimizers can be used to improve the performance of surrogate models for structural optimization under uncert…
Source: AegisMind Research
Read full discoveryPhysics + Computer ScienceMar 17, 2026Evaluation Score: 53%
Amortized surrogates for mRNA design can be trained using the soft feasibility enforcement strategy from Cheap Thrills t…
Source: AegisMind Research
Read full discoveryPhysics + Computer ScienceMar 17, 2026Evaluation Score: 53%
The projection-based model order reduction used in structural optimization can be adapted to compress the state space of…
Source: AegisMind Research
Read full discoveryPhysics + Computer ScienceMar 17, 2026Evaluation Score: 53%
Uncertainty quantification methods from reduced-order structural models can be integrated into FlashOptim's mixed-precis…
Source: AegisMind Research
Read full discoveryPhysics + Computer ScienceMar 17, 2026Evaluation Score: 53%
Cheap surrogate models can accelerate the optimization of mRNA sequences for improved protein production.
Source: AegisMind Research
Read full discoveryPhysics + Computer ScienceMar 17, 2026Evaluation Score: 53%
Low-rank approximation can reduce the computational cost of training LLMs for financial trading.
Source: AegisMind Research
Read full discoveryPhysics + Computer ScienceMar 17, 2026Evaluation Score: 53%
FlashOptim techniques can reduce memory requirements for training models that predict mRNA stability.
Source: AegisMind Research
Read full discoveryPhysics + Computer ScienceMar 17, 2026Evaluation Score: 55%
Uncertainty-aware gradients from matrix-interpolatory ROMs will enhance adaptive sampling for mRNA stability optimizatio…
Source: AegisMind Research
Read full discoveryPhysics + Computer ScienceMar 17, 2026Evaluation Score: 55%
Taming Momentum's EMA reframing as low-rank matrix updates can be theor
Source: AegisMind Research
Read full discoveryPhysics + Computer ScienceMar 17, 2026Evaluation Score: 55%
Adaptive sampling algorithms can improve the efficiency of training LLMs for investment by focusing on informative marke…
Source: AegisMind Research
Read full discoveryPhysics + Computer ScienceMar 17, 2026Evaluation Score: 55%
FlashOptim techniques can reduce the memory footprint of training models for predicting confluent tissue dynamics.
Source: AegisMind Research
Read full discoveryPhysics + Computer ScienceMar 17, 2026Evaluation Score: 55%
Adaptive sampling algorithms can be used to optimize the parameters of active foam models, improving simulation accuracy…
Source: AegisMind Research
Read full discoveryPhysics + Computer ScienceMar 17, 2026Evaluation Score: 55%
Applying low-rank approximation to optimizer states in LLM investment agents will reduce memory overhead and improve tra…
Source: AegisMind Research
Read full discoveryPhysics + Computer ScienceMar 17, 2026Evaluation Score: 57%
Taming Momentum approximations will reduce optimizer states in multi-agent trading LLMs, enabling larger team simulation…
Source: AegisMind Research
Read full discoveryPhysics + Computer ScienceMar 17, 2026Evaluation Score: 57%
FlashOptim's memory-efficient training scheme can enable larger surrogate networks for amortized optimization without ex…
Source: AegisMind Research
Read full discoveryPhysics + Computer ScienceMar 17, 2026Evaluation Score: 57%
Low-rank approximation of optimizer states can improve the scalability of training surrogate models for structural optim…
Source: AegisMind Research
Read full discoveryPhysics + Computer ScienceMar 17, 2026Evaluation Score: 60%
Multi-agent LLM trading systems can incorporate amortized optimization surrogates to replace expensive portfolio simulat…
Source: AegisMind Research
Read full discoveryPhysics + Computer ScienceMar 17, 2026Evaluation Score: 60%
Low-rank approximation of optimizer momentum states (as in Taming Momentum) can be applied to evolutionary LLM-driven op…
Source: AegisMind Research
Read full discoveryPhysics + Computer ScienceMar 17, 2026Evaluation Score: 65%
Low-rank EMA approximations of optimizer states can reduce memory consumption in training surrogate models for structura…
Source: AegisMind Research
Read full discoveryPhysics + Computer ScienceMar 17, 2026Evaluation Score: 65%
The random-key optimizer framework for MIPs can be augmented with LLM-generated semantic mutations analogous to AdaEvolv…
Source: AegisMind Research
Read full discoveryPhysics + Computer ScienceMar 17, 2026Evaluation Score: 65%
FlashOptim techniques can reduce memory requirements for training LLM-powered investment agents, enabling larger models.
Source: AegisMind Research
Read full discoveryComputer Science + PhysicsMar 12, 2026Evaluation Score: 45%
Amortized optimization surrogates trained on inexpensive labels can accelerate mRNA sequence design by replacing costly …
Source: AegisMind Research
Read full discoveryComputer Science + PhysicsMar 12, 2026Evaluation Score: 47%
Amortized optimization with cheap labels can approximate surrogate fitness functions for evolutionary mRNA design, reduc…
Source: AegisMind Research
Read full discoveryComputer Science + PhysicsMar 12, 2026Evaluation Score: 53%
Cheap-label amortized optimization can train surrogate models for mixed-integer program feasibility prediction, accelera…
Source: AegisMind Research
Read full discoveryComputer Science + PhysicsMar 12, 2026Evaluation Score: 60%
Adaptive sampling strategies from uncertainty-aware structural optimization can improve exploration efficiency in LLM-dr…
Source: AegisMind Research
Read full discoveryComputer Science + PhysicsMar 12, 2026Evaluation Score: 57%
Amortized optimization with inexpensive labels can generate approximate warm-start solutions for MIP solvers, reducing b…
Source: AegisMind Research
Read full discoveryComputer Science + PhysicsMar 12, 2026Evaluation Score: 60%
Low-rank momentum approximation reduces memory sufficiently to enable on-device fine-tuning of LLMs used as semantic mut…
Source: AegisMind Research
Read full discoveryComputer Science + PhysicsMar 12, 2026Evaluation Score: 63%
Model order reduction techniques for structural optimization can accelerate the simulation backbone of amortized optimiz…
Source: AegisMind Research
Read full discoveryComputer Science + PhysicsMar 12, 2026Evaluation Score: 67%
Adaptive LLM-driven zeroth-order optimization schedules can dynamically adjust mutation rates in mRNA sequence design ev…
Source: AegisMind Research
Read full discoveryComputer Science + PhysicsMar 12, 2026Evaluation Score: 57%
Uncertainty-aware adaptive sampling from projection-based reduced-order models can improve the efficiency of amortized o…
Source: AegisMind Research
Read full discoveryPhysics + Computer ScienceMar 12, 2026Evaluation Score: 47%
FlashOptim's memory-efficient training strategies can enable on-device fine-tuning of LLMs used as mutation operators in…
Source: AegisMind Research
Read full discoveryPhysics + Computer ScienceMar 12, 2026Evaluation Score: 47%
Amortized optimization networks trained on inexpensive labels can accelerate mRNA sequence design by learning surrogate …
Source: AegisMind Research
Read full discoveryPhysics + Computer ScienceMar 12, 2026Evaluation Score: 50%
Low-rank EMA representations from Taming Momentum can compress the optimizer state in mRNA sequence optimization, enabli…
Source: AegisMind Research
Read full discoveryPhysics + Computer ScienceMar 12, 2026Evaluation Score: 50%
Uncertainty-aware gradient calculations from reduced-order structural models can be integrated into amortized optimizati…
Source: AegisMind Research
Read full discoveryPhysics + Computer ScienceMar 12, 2026Evaluation Score: 50%
Adaptive sampling strategies from uncertainty-aware reduced-order models can improve the efficiency of sampling-based co…
Source: AegisMind Research
Read full discoveryPhysics + Computer ScienceMar 12, 2026Evaluation Score: 57%
Cheap-label amortized optimization can reduce the computational cost of evaluating candidate solutions in random-key opt…
Source: AegisMind Research
Read full discoveryPhysics + Computer ScienceMar 12, 2026Evaluation Score: 57%
Low-rank approximation of optimizer momentum states (as in Taming Momentum) can reduce memory overhead in training LLM-b…
Source: AegisMind Research
Read full discoveryComputer Science + PhysicsMar 11, 2026Evaluation Score: 47%
The low-rank EMA reframing in Taming Momentum can be applied to optimizer states in amortized optimization networks, red…
Source: AegisMind Research
Read full discoveryComputer Science + PhysicsMar 11, 2026Evaluation Score: 50%
Random-key optimizer frameworks applied to mixed-integer programs can be enhanced with LLM-generated heuristics from Ada…
Source: AegisMind Research
Read full discoveryComputer Science + PhysicsMar 11, 2026Evaluation Score: 57%
FlashOptim's memory compression strategies can be combined with low-rank momentum approximation from Taming Momentum to …
Source: AegisMind Research
Read full discoveryComputer Science + PhysicsMar 11, 2026Evaluation Score: 50%
Uncertainty-aware gradient calculations for reduced-order structural models can be adapted to quantify prediction uncert…
Source: AegisMind Research
Read full discoveryComputer Science + PhysicsMar 11, 2026Evaluation Score: 50%
Adaptive LLM-driven zeroth-order optimization can replace gradient-based search in mRNA synonymous space by treating cod…
Source: AegisMind Research
Read full discoveryComputer Science + PhysicsMar 11, 2026Evaluation Score: 57%
FlashOptim's memory-efficient optimizer states can enable training of larger surrogate networks for amortized optimizati…
Source: AegisMind Research
Read full discoveryComputer Science + PhysicsMar 11, 2026Evaluation Score: 67%
The evolutionary loop in AdaEvolve can incorporate a reduced-order model of the fitness landscape, analogous to structur…
EVP availableSource: AegisMind Research
Read full discoveryPhysics + Computer ScienceMar 11, 2026Evaluation Score: 47%
Amortized optimization using ML surrogates can accelerate the design of synthetic microbial consortia by predicting opti…
Source: AegisMind Research
Read full discoveryPhysics + Computer ScienceMar 11, 2026Evaluation Score: 50%
AdaEvolve can be used to optimize the objective functions used in mRNA sequence design, leading to improved performance.
Source: AegisMind Research
Read full discoveryPhysics + Computer ScienceMar 11, 2026Evaluation Score: 53%
Uncertainty-aware adaptive sampling in reduced-order models can be reformulated as an active learning strategy for train…
Source: AegisMind Research
Read full discoveryPhysics + Computer ScienceMar 11, 2026Evaluation Score: 53%
Cheap Thrills methods using ML surrogates can accelerate parameter tuning in agent-based models of confluent tissue dyna…
Source: AegisMind Research
Read full discoveryPhysics + Computer ScienceMar 11, 2026Evaluation Score: 53%
LLMs can serve as effective surrogates for computationally expensive steps in structural optimization, accelerating the …
Source: AegisMind Research
Read full discoveryPhysics + Computer ScienceMar 11, 2026Evaluation Score: 63%
Cheap inexpensive-label surrogates can replace expensive finite-element evaluations in structural optimization by learni…
Source: AegisMind Research
Read full discoveryPhysics + Computer ScienceMar 11, 2026Evaluation Score: 47%
mRNA design optimization using sampling-based continuous methods can benefit from low-rank approximations of the codon c…
Source: AegisMind Research
Read full discoveryPhysics + Computer ScienceMar 11, 2026Evaluation Score: 47%
Low-rank approximation of optimizer states can improve the scalability of training LLMs for modeling confluent tissue be…
Source: AegisMind Research
Read full discoveryPhysics + Computer ScienceMar 11, 2026Evaluation Score: 50%
Low-rank approximation can reduce the memory requirements of training models that predict the behavior of synthetic micr…
Source: AegisMind Research
Read full discoveryPhysics + Computer ScienceMar 11, 2026Evaluation Score: 50%
Applying low-rank approximation to optimizer states in LLM-driven investment agents will reduce memory overhead and impr…
Source: AegisMind Research
Read full discoveryPhysics + Computer ScienceMar 11, 2026Evaluation Score: 53%
FlashOptim's quantized optimizer states can be applied to reduce memory in training surrogate models for large-scale str…
Source: AegisMind Research
Read full discoveryPhysics + Computer ScienceMar 11, 2026Evaluation Score: 57%
AdaEvolve's adaptive LLM mutation scheduling can improve mixed-integer program solving by dynamically adjusting explorat…
Source: AegisMind Research
Read full discoveryPhysics + Computer ScienceMar 11, 2026Evaluation Score: 57%
Low-rank approximation of optimizer momentum states (as in Taming Momentum) can reduce memory overhead in training LLM-b…
Source: AegisMind Research
Read full discovery