Object rearrangement is a challenge for embodied agents because solving these tasks requires generalizing across a combinatorially large set of configurations of entities and their locations. Worse, ...the representations of these entities are unknown and must be inferred from sensory percepts. We present a hierarchical abstraction approach to uncover these underlying entities and achieve combinatorial generalization from unstructured visual inputs. By constructing a factorized transition graph over clusters of entity representations inferred from pixels, we show how to learn a correspondence between intervening on states of entities in the agent's model and acting on objects in the environment. We use this correspondence to develop a method for control that generalizes to different numbers and configurations of objects, which outperforms current offline deep RL methods when evaluated on simulated rearrangement tasks.
DBA/1 mice develop a chronic peripheral arthritis after immunization with type II collagen termed collagen-induced arthritis. We have localized the main arthritogenic determinants of CB11, a ...CNBr-generated arthritogenic fragment of chick type II collagen (CII), using 3 smaller peptide fragments of CB11 generated by endoproteinase LysC, LysC1 (CII 124-290), LysC2 (CII 291-374) and LysC3 (CII 375-402) and a panel of monoclonal antibodies (mAb) specific to CB11. MAb specific to the arthritogenic region of CB11 were also used to study the synergistic effect of E. coli lipopolysaccharide (LPS) on antibody-mediated arthritis in naive DBA/1 mice.
LysC2 contained a minimum essential arthritogenic fragment of type II collagen: LysC2 induced arthritis by active immunization, also, a combination of four mAb specific to LysC2 passively transferred arthritis to naive mice. A single i.p. injection of LPS (50 μg/mouse) reduced the threshold values of the arthritogenic dose of mAb from 1 mg to 50 μg/clone per mouse, and decreased the number of mAb required for inducing arthritis from 4 to 2 clones. These observations suggest that LysC2, an 84 amino acid residue fragment, contains the main arthritogenic determinants within chick CB11. Importantly, LPS, a strong inducer of pro-inflammatory cytokines, negates the required multiple epitope specificity of autoantibodies in the passive transfer model and acts synergistically in the induction of arthritis by autoantibody.
Kidney transplantation is the best treatment for patients with end-stage renal failure, but uncertainty remains about the best immunosuppression strategy. Long-term graft survival has not improved ...substantially, and one possible explanation is calcineurin inhibitor (CNI) nephrotoxicity. CNI exposure could be minimized by using more potent induction therapy or alternative maintenance therapy to remove CNIs completely. However, the safety and efficacy of such strategies are unknown.
The Campath, Calcineurin inhibitor reduction and Chronic allograft nephropathy (3C) Study is a multicentre, open-label, randomized controlled trial with 852 participants which is addressing two important questions in kidney transplantation. The first question is whether a Campath (alemtuzumab)-based induction therapy strategy is superior to basiliximab-based therapy, and the second is whether, from 6 months after transplantation, a sirolimus-based maintenance therapy strategy is superior to tacrolimus-based therapy. Recruitment is complete, and follow-up will continue for around 5 years post-transplant. The primary endpoint for the induction therapy comparison is biopsy-proven acute rejection by 6 months, and the primary endpoint for the maintenance therapy comparison is change in estimated glomerular filtration rate from baseline to 2 years after transplantation. The study is sponsored by the University of Oxford and endorsed by the British Transplantation Society, and 18 centers for adult kidney transplant are participating.
Late graft failure is a major issue for kidney-transplant recipients. If our hypothesis that minimizing CNI exposure with Campath-based induction therapy and/or an elective conversion to sirolimus-based maintenance therapy can improve long-term graft function and survival is correct, then patients should experience better graft function for longer. A positive outcome could change clinical practice in kidney transplantation.
ClinicalTrials.gov, NCT01120028 and ISRCTN88894088.
Large language models are few-shot learners that can solve diverse tasks from a handful of demonstrations. This implicit understanding of tasks suggests that the attention mechanisms over word tokens ...may play a role in analogical reasoning. In this work, we investigate whether analogical reasoning can enable in-context composition over composable elements of visual stimuli. First, we introduce a suite of three benchmarks to test the generalization properties of a visual in-context learner. We formalize the notion of an analogy-based in-context learner and use it to design a meta-learning framework called Im-Promptu. Whereas the requisite token granularity for language is well established, the appropriate compositional granularity for enabling in-context generalization in visual stimuli is usually unspecified. To this end, we use Im-Promptu to train multiple agents with different levels of compositionality, including vector representations, patch representations, and object slots. Our experiments reveal tradeoffs between extrapolation abilities and the degree of compositionality, with non-compositional representations extending learned composition rules to unseen domains but performing poorly on combinatorial tasks. Patch-based representations require patches to contain entire objects for robust extrapolation. At the same time, object-centric tokenizers coupled with a cross-attention module generate consistent and high-fidelity solutions, with these inductive biases being particularly crucial for compositional generalization. Lastly, we demonstrate a use case of Im-Promptu as an intuitive programming interface for image generation.
Multiple Sclerosis (MS) is a chronic neurological disease characterized by demyelination associated with infiltrating white blood cells in the central nervous system (CNS). Nitric oxide synthases ...(NOS) are a family of enzymes that control the production of nitric oxide. It is possible that neuronal NOS could be involved in MS pathophysiology and hence the nNOS gene is a potential candidate for involvement in disease susceptibility. The aim of this study was to determine whether allelic variation at the nNOS gene locus is associated with MS in an Australian cohort. DNA samples obtained from a Caucasian Australian population affected with MS and an unaffected control population, matched for gender, age and ethnicity, were genotyped for a microsatellite polymorphism in the promoter region of the nNOS gene. Allele frequencies were compared using chi-squared based statistical analyses with significance tested by Monte Carlo simulation. Allelic analysis of MS cases and controls produced a chi-squared value of 5.63 with simulated
P=0.96 (OR
(max)=1.41, 95% CI: 0.926–2.15). Similarly, a Mann–Whitney
U analysis gave a non-significant
P-value of 0.377 for allele distribution. No differences in allele frequencies were observed for gender or clinical course subtype (
P>0.05). Statistical analysis indicated that there is no association of this nNOS variant and MS and hence the gene does not appear to play a genetically significant role in disease susceptibility.
Full text
Available for:
GEOZS, IJS, IMTLJ, KILJ, KISLJ, NUK, OILJ, PNG, SAZU, SBCE, SBJE, UL, UM, UPCLJ, UPUK
Antibodies binding to double‐stranded (ds) DNA are strongly associated with renal involvement in patients with systemic lupus erythematosus (SLE). We have generated two new IgG DNA‐binding monoclonal ...antibodies (mAb), RH‐14 and DIL‐6, from the peripheral blood lymphocytes of two SLE patients with glomerulonephritis using the heteromyeloma cell line CB‐F7. RH‐14 is an IgG1 λ antibody which also bound to single‐stranded (ss)DNA, histones and nucleosomes. DIL‐6 is an IgG3 λ antibody with restricted antigen binding specificity. cDNA encoding the variable regions of the heavy (VH) and light (VL) chains of RH‐14 was sequenced and the antigen binding site of this mAb was computer modelled. Sequence analysis of VH and VL regions of RH‐14 showed that VH is derived from germ‐line gene V3‐7, a member of the VH3 family, and VL is derived from DPL 11, a member of the Vλ2 family. Somatic mutations and basic amino acid residues are identified in the complementarity‐ determining regions of both VH and VL regions. The nephritogenic properties of these mAb were analyzed by implanting and growing the hybridoma cells secreting the mAb in the peri toneum of SCID mice. The animals that received the RH‐14 hybridoma produced higher levels of proteinuria (3 to ≥ 4) (p < 0.001) compared to the groups that received DIL‐6 (trace to ≥ 1) or CB‐F7 (trace). Electron microscopy of kidney sections from all the RH‐14‐ implanted animals showed granular immunoglobulin deposition in the renal glomerular capil laries and mesangium. In this study we have shown for the first time using electron micros copy that a human IgG anti‐dsDNA mAb, RH‐14, is nephritogenic and that deposition of such an antibody alone is sufficient to induce renal damage.
Full text
Available for:
FZAB, GIS, IJS, KILJ, NLZOH, NUK, OILJ, SBCE, SBMB, UL, UM, UPUK
Many transfer problems require re-using previously optimal decisions for solving new tasks, which suggests the need for learning algorithms that can modify the mechanisms for choosing certain actions ...independently of those for choosing others. However, there is currently no formalism nor theory for how to achieve this kind of modular credit assignment. To answer this question, we define modular credit assignment as a constraint on minimizing the algorithmic mutual information among feedback signals for different decisions. We introduce what we call the modularity criterion for testing whether a learning algorithm satisfies this constraint by performing causal analysis on the algorithm itself. We generalize the recently proposed societal decision-making framework as a more granular formalism than the Markov decision process to prove that for decision sequences that do not contain cycles, certain single-step temporal difference action-value methods meet this criterion while all policy-gradient methods do not. Empirical evidence suggests that such action-value methods are more sample efficient than policy-gradient methods on transfer problems that require only sparse changes to a sequence of previously optimal decisions.
The cytoskeletal proteins of cultured normal human mesothelial cells were found to consist of six major components, including actin, vimentin, the 40 kd keratin and the 44, 52 and 55 kd proteins, ...plus a minor 46 kd protein. Two-dimensional gel electrophoresis, peptide mapping and immunoprecipitation tests showed that the 40-55 kd mesothelial proteins are a family of keratins distinct in size, charge or peptide map from the "epidermal keratins" synthesized by cultured keratinocytes. Unique combinations of keratins from the epidermal and mesothelial keratin families were found to be synthesized by cultured bladder, esophageal, conjunctival, mammary, exocervical and ovarian surface epithelial cells. Mesothelial cells were the only epithelial cell type that synthesized vimentin at more than trace levels. We have also found that many carcinoma cell lines express keratins different from those of their cell type of origin.
Humans extract useful abstractions of the world from noisy sensory data. Serial reproduction allows us to study how people construe the world through a paradigm similar to the game of telephone, ...where one person observes a stimulus and reproduces it for the next to form a chain of reproductions. Past serial reproduction experiments typically employ a single sensory modality, but humans often communicate abstractions of the world to each other through language. To investigate the effect language on the formation of abstractions, we implement a novel multimodal serial reproduction framework by asking people who receive a visual stimulus to reproduce it in a linguistic format, and vice versa. We ran unimodal and multimodal chains with both humans and GPT-4 and find that adding language as a modality has a larger effect on human reproductions than GPT-4's. This suggests human visual and linguistic representations are more dissociable than those of GPT-4.